Overview of NASA Earth Observing Systems Terra and Aqua ...€¦ · [email protected] cJoint...

Overview of NASA Earth Observing Systems Terra and Aqua moderate resolution imaging

spectroradiometer instrument calibration algorithms and on-orbit performance

Xiaoxiong (Jack) Xiong,a Brian N. Wenny,b and William L. Barnesc

aSciences and Exploration Directorate, NASA/GSFC, Greenbelt, MD 20771 [email protected]

bScience Systems and Applications, Inc., 10210 Greenbelt Road, Lanham, MD 20706 [email protected]

cJoint Center for Earth Systems Technology/University of Maryland Baltimore County, 5523 Research Park Drive, Baltimore, MD 21228

[email protected]

Abstract. Since launch, the Terra and Aqua moderate resolution imaging spectroradiometer (MODIS) instruments have successfully operated on-orbit for more than 9 and 6.5 years, respectively. MODIS, a key instrument for the NASA’s Earth Observing System (EOS) missions, was designed to make continuous observations for studies of Earth’s land, ocean, and atmospheric properties and to extend existing data records from heritage earth-observing sensors. In addition to frequent global coverage, MODIS observations are made in 36 spectral bands, covering both solar reflective and thermal emissive spectral regions. Nearly 40 data products are routinely generated from MODIS observations and publicly distributed for a broad range of applications. Both instruments have produced an unprecedented amount of data in support of the science community. As a general reference for understanding sensor operation and calibration, and thus science data quality, we provide an overview of the MODIS instruments and their pre-launch calibration and characterization, and describe their on-orbit calibration algorithms and performance. On-orbit results from both Terra and Aqua MODIS radiometric, spectral, and spatial calibration are discussed. Currently, both instruments, including their on-board calibration devices, are healthy and are expected to continue operation for several years to come. Keywords: MODIS, calibration, radiometric, spectral, spatial, on-board calibrators.

1 INTRODUCTION The NASA’s Moderate Resolution Imaging Spectroradiometer (MODIS) was designed and developed to collect continuous global data for studies of both short- and long-term changes in the Earth’s land, ocean and atmosphere systems and to help the science community assess the impact of global environmental and climate changes. One of the design objectives of the MODIS instruments was to continue and enhance the existing data records produced by heritage sensors such as the AVHRR, Landsat TM, CZCS, and HIRS [see Appendix A for acronym definitions] with an overall improvement of its spatial, spectral, and temporal characteristics [1-3]. Because of its capability to address a broad range of interdisciplinary science objectives with frequent multi-spectral global observations, MODIS has been described as the “keystone” instrument for both the Terra and Aqua missions. The Terra spacecraft was launched in December 1999, carrying five EOS sensors: ASTER, CERES, MISR, MOPITT, and MODIS. The Aqua spacecraft was launched in May 2002, providing a platform for six EOS sensors: AIRS, AMSR-E, AMSU, CERES, HSB, and MODIS. Both spacecraft are operated in near sun-synchronous polar orbits at a nominal altitude of 705 km,

Journal of Applied Remote Sensing, Vol. 3, 032501 (26 June 2009)

© 2009 Society of Photo-Optical Instrumentation Engineers [DOI: 10.1117/1.3180864]Received 11 Jun 2009; accepted 24 Jun 2009; published 26 Jun 2009 [CCC: 19313195/2009/$25.00]Journal of Applied Remote Sensing, Vol. 3, 032501 (2009)

Downloaded From: http://remotesensing.spiedigitallibrary.org/ on 03/07/2016 Terms of Use: http://spiedigitallibrary.org/ss/TermsOfUse.aspx

mailto:[email protected]




thus enabling complementary morning (Terra) and afternoon (Aqua) observations. Terra, with an equatorial crossing time of 10:30 LST, is part of the AM constellation, which also includes Landsat-7 and EO-1. Aqua has an equatorial crossing time of 13:30 LST. It is part of the PM constellation, also known as the “A-Train”, which includes the Aura, PARASOL, CALIPSO, and CloudSat satellite missions. The first MODIS instrument, the Protoflight Model (PFM), is on-board the Terra spacecraft and is thus referred to as the Terra MODIS. The second MODIS instrument, the Flight Model 1 (FM1), is referred to as the Aqua MODIS. The two MODIS instruments, built with the same design specifications, are nearly identical. Each MODIS instrument has 20 reflective solar bands (RSB) and 16 thermal emissive bands (TEB), with wavelengths from 0.41 to 2.2µm and 3.75 to 14.24µm, respectively. MODIS observations are made at three nadir spatial resolutions: 250m for bands 1-2, 500m for bands 3-7, and 1km for bands 8-36. Table 1 is a summary of key design parameters and calibration requirements of the MODIS spectral bands, including their primary science applications. One of the most significant improvements of the MODIS instrument over heritage sensors is its advanced capability of performing comprehensive on-orbit calibration and characterization. This is achieved through a set of on-board calibrators (OBC), as shown in Figure 1.

Fig. 1. MODIS instrument cavity showing the location of key on-board calibrators and optical components. Since launch, Terra MODIS has successfully operated for more than 9 years and Aqua MODIS for nearly 7 years. Together, they have produced an unprecedented amount of high-quality data for the science community and significantly contributed to satellite remote sensing studies and applications. Nearly 40 science data products are routinely produced and distributed worldwide in support of scientific studies of a broad range of land, ocean and atmospheric properties [4-8]. Each science discipline, which can be accessed via the MODIS URL at modis.gsfc.nasa.gov, provides information on their respective products. All of the data products are available to the global science community (modis.gsfc.nasa.gov/data/). To date, both instruments are in a healthy operational condition. Their on-board calibrators continue to function normally, allowing calibration and data quality to be maintained. Considerable effort has been made by the MODIS Characterization Support Team (MCST) at NASA/GSFC to assure satisfactory operation and calibration for the extended missions of both Terra and Aqua MODIS and to continue producing high quality data products. Updates of instrument status, information on the MODIS Level 1B data products, and changes in the

Journal of Applied Remote Sensing, Vol. 3, 032501 (2009)


http://modis.gsfc.nasa.gov/

http://modis.gsfc.nasa.gov/data/

calibration algorithms and look-up tables (LUT) are provided through the MCST web page at www.mcst.ssai.biz/mcstweb/index.html. Table 1. MODIS spectral band design specifications. CW is center wavelength in µm; BW is bandwidth in µm; Ltyp is typical radiance in W/m2/sr/µm; Ttyp is typical temperature in K; SNR is signal-to-noise ratio; NEdT is noise equivalent difference temperature in K.

RSB Band CW BW Ltyp SNR Primary Use 1 0.645 0.050 21.8 128 2 0.858 0.035 24.7 201 Land/cloud/aerosols boundaries

3 0.469 0.020 35.3 243 4 0.555 0.020 29.0 228 5 1.240 0.020 5.4 74 6 1.640 0.024 7.3 275 7 2.130 0.050 1.0 110

Land/cloud/aerosols properties

8 0.412 0.015 44.9 880 9 0.443 0.010 41.9 838

10 0.488 0.010 32.1 802 11 0.531 0.010 27.9 754 12 0.551 0.010 21.0 750 13 0.667 0.010 9.5 910 14 0.678 0.010 8.7 1087 15 0.748 0.010 10.2 586 16 0.869 0.015 6.2 516

Ocean color, phytoplankton & biogeochemistry

17 0.905 0.030 10.0 167 18 0.936 0.010 3.6 57 19 0.940 0.050 15.0 250

Atmospheric water vapor

26 1.375 0.030 6.0 150 Cirrus clouds water vapor TEB Band CW BW Ttyp NEdT Primary Use

20 3.75 0.18 300 0.05 21 3.96 0.06 335 0.20 22 3.96 0.06 300 0.07 23 4.05 0.06 300 0.07

Surface/cloud temperature

24 4.47 0.07 250 0.25 25 4.52 0.07 275 0.25 Atmospheric temperature

27 6.72 0.36 240 0.25 28 7.33 0.30 250 0.25 Water vapor

29 8.55 0.30 300 0.05 Cloud properties 30 9.73 0.30 250 0.25 Ozone 31 11.03 0.50 300 0.05 32 12.02 0.50 300 0.05 Surface/cloud temperature

33 13.34 0.30 260 0.25 34 13.64 0.30 250 0.25 35 13.94 0.30 240 0.25 36 14.24 0.30 220 0.35

Cloud top altitude

The objective of this paper is to provide an overview of both Terra and Aqua MODIS calibration and performance. This review is directed to the general remote sensing community in order to facilitate a better understanding of sensor calibration and characterization and thus the impact of MODIS instrument performance on the generated science products. Section 2 provides a brief description of MODIS, including its key components and sub-systems, and an



http://www.mcst.ssai.biz/mcstweb/index.html

overview of its on-board calibration systems. Pre-launch calibration activities are presented in Section 3. Section 4 describes instrument on-orbit calibration methodologies and algorithms. In Section 5, on-orbit results are illustrated and discussed for both Terra and Aqua MODIS. Additional discussions focusing on lessons learned and applications to future missions are included in Section 6. A summary of this paper is given in Section 7. Also included is an appendix of acronyms.

2 INSTRUMENT BACKGROUND MODIS is a multispectral filter radiometer, making observations in 36 discrete spectral bands. It collects data using both sides of its scan mirror. The radiant flux reflected from the scan mirror is directed by a fold mirror to the sensor’s off-axis telescope, consisting of a primary and a secondary mirror as illustrated in Figure 1. The aft optics includes 3 beam splitters, 4 objective assemblies, and various blocking and spectral band pass filters. The 36 spectral bands are located on four focal plane assemblies (FPA), shown in Figure 2, according to their spectral regions: visible (VIS), near infrared (NIR), short- and mid-wave infrared (SMIR), and long-wave infrared (LWIR). The temperatures of the VIS and NIR FPA are not controlled, and float with the instrument, whereas the SMIR and LWIR FPA are controlled nominally at 83K via a passive radiative cooler and are referred to as the cold FPA (CFPA).

Fig. 2. Bands and detectors on the four MODIS focal plane assemblies. The scan and track direction, optical axis and detector type for each FPA are included. There are 40, 20, and 10 detectors for each 250m, 500m, and 1km resolution spectral band, respectively, with the exception of bands 13 and 14, which make both high and low gain observations through a time-delay and integration (TDI) with a pair of 10-detector arrays. As a result, MODIS has a total of 490 detectors. The spectral bands are aligned in the along-scan direction and detectors in each spectral band in the along-track direction. Both sides of the scan mirror collect data alternately. The Earth view (EV) observations are made over a scan angle range of ±55º relative to instrument nadir, resulting in a swath of 2330km cross-track by 10 km (at nadir) along-track in each scan of 1.478s and, therefore, producing a complete global coverage in less than 2 days. To meet its on-orbit calibration requirements, MODIS

LWIR

30 291

2

3

4

5

6

7

8

9

10

28 271

2

3

4

5

6

7

8

9

10

331

2

3

4

5

6

7

8

9

10

341

2

3

4

5

6

7

8

9

10

351

2

3

4

5

6

7

8

9

10

361

2

3

4

5

6

7

8

9

10

311

2

3

4

5

6

7

8

9

10

32

251

2

3

4

5

6

7

8

9

10

241

2

3

4

5

6

7

8

9

10

26 7123

4

5

67891011

121314

151617181920

6123

4

5

67891011

121314

151617181920

5123

45

67891011

121314

1516171819

20

20 21 22 231

2

3

4

5

6

7

8

9

10

SWIR/MWIRFILTERS

DETECTORS

10 9 8 3123456

78

9101112131415

1617

181920

412345

6

7

89101112131415

16

17181920

11 121

2

3

4

5

6

7

8

9

10

VIS

NIR

FILTERS

DETECTORS

171

2

3

4

5

6

7

8

9

10

16151414'11

40

21313'18191

2

3

4

5

6

7

8

9

10

S

T

1

40

S

T

S

T

S

T

Silicon

PV HgCdTe PC HgCdTeSilicon

PV HgCdTe

= OPTICAL AXIS

1

2

3

4

5

6

7

8

9

10

1

2

3

4

5

6

7

8

9

10

1

2

3

4

5

6

7

8

9

10

1

2

3

4

5

6

7

8

9

10

1

2

3

4

5

6

7

8

9

10

1

2

3

4

5

6

7

8

9

10

1

2

3

4

5

6

7

8

9

10

1

2

3

4

5

6

7

8

9

10

1

2

3

4

5

6

7

8

9

10

1

2

3

4

5

6

7

8

9

10

1

2

3

4

5

6

7

8

9

10

1

2

3

4

5

6

7

8

9

10

1

2

3

4

5

6

7

8

9

10

1

2

3

4

5

6

7

8

9

10

1

2

3

4

5

6

7

8

9

10

1

2

3

4

5

6

7

8

9

10

1

2

3

4

5

6

7

8

9

10

1

2

3

4

5

6

7

8

9

10



was built with an extensive set of on-board calibrators (OBC) capable of performing different calibration functions. These include a solar diffuser (SD), a solar diffuser stability monitor (SDSM), a blackbody (BB), and a spectro-radiometric calibration assembly (SRCA). A space view (SV) port is often considered part of the OBC. Observations through the SV port provide measurements of instrument background for a zero signal reference. The SD panel, made of space-grade Spectralon, is used as the RSB on-orbit calibration reference. The SDSM is designed to operate during SD calibrations in order to track its on-orbit degradation. As a ratioing radiometer, the SDSM makes alternate measurements of direct sunlight through an attenuation screen and the sunlight reflected from the SD at 9 wavelengths from 0.41 to 0.96µm. The attenuation screen was designed to match the signals from the SDSM Sun view and SD view. The relative change in the time-dependent ratios of the SD view responses to the Sun view responses is a measure of the SD degradation. The on-board BB, with its temperatures measured by a set of 12 thermistors, provides a calibration reference for the TEB. The SRCA is a unique device with built-in sources and a limited internal spectral calibration capability. It can be configured and operated in three modes: radiometric, spectral, and spatial. The SRCA radiometric and spectral calibration capabilities only apply to the RSB, while the spatial calibration includes both the TEB and RSB [9-10]. For each 1km resolution detector, MODIS collects 50, 10, 50, 50, and 1354 frames (data samples) per scan over the SD, SRCA, BB, SV, and EV sectors, respectively. The number of data samples are doubled for each 500m resolution detector and quadrupled for each 250m resolution detector.

3 PRE-LAUNCH CALIBRATION ACTIVITIES MODIS development was driven by the desire of a broad science community for a number of improvements over its heritage sensors, including more stringent calibration requirements and improved on-orbit calibration. Consequently, extensive pre-launch measurements were made to calibrate and characterize sensor performance. Described in the following are the key radiometric, spectral, and spatial calibration and characterization activities performed pre-launch, and the resulting issues. MODIS pre-launch radiometric calibration included measurements of each detector’s gain, nonlinearity, dynamic range, signal-to-noise ratio (SNR), and response repeatability. For the purpose of setting each detector’s gain and offset, limited radiometric calibration was initially performed at component and sub-system levels in ambient conditions. After sensor integration and test (I&T), a complete set of system-level radiometric calibrations and characterizations was made in a thermal vacuum chamber in order to fully evaluate sensor performance and design compliance. Figure 3 shows the layout of MODIS in the thermal vacuum chamber. In anticipation of sensor launch, activation, and on-orbit operational conditions at the beginning of life (BOL) and end of life (EOL), key radiometric calibrations were performed at three instrument temperatures, often referred to as the cold, nominal, and hot plateaus. As an enhancement to the sensor reliability, MODIS was built with primary (A-side) and redundant (B-side) electronic subsystems, including the power supply, time-generator, control processor (analog), and formatter (digital). For Terra MODIS, both primary and redundant configurations were characterized in the thermal vacuum at instrument nominal plateau. With lessons learned and issues uncovered from Terra MODIS, additional configurations, including a number of cross-strapping configurations, were also characterized for Aqua MODIS, which led to an improved understanding of its performance. A spherical integrating source (SIS) was used as the primary calibration source for the RSB radiometric calibration. The SIS was operated in an ambient environment. It was characterized with traceability to NIST irradiance standards. By operating the SIS with different lamp configurations, multiple radiance levels are generated. The RSB detector gain (or 1/response), nonlinearity, dynamic range, and signal-to-noise ratio (SNR) were determined



from detector responses to multiple input radiances from the SIS. The repeatability was characterized using measurements made under the same test condition but at different times. The thermal emissive bands (TEB) radiometric calibration was carried out using a ground-based blackbody calibration source (BCS), with a modeled emissivity of better than 0.9998 across the entire TEB spectral range. The temperatures of the BCS were measured using high-precision thermistors. All thermistors were fully characterized with traceability to NIST temperature scales. Different radiance levels were achieved by operating the BCS at different temperatures. The BCS was operated within the thermal vacuum chamber, allowing the sensor to view the calibration target directly. In addition to characterizations made at different instrument temperatures, the TEB radiometric calibration was made at three different CFPA temperatures, including the nominal temperature of 83K. The on-board BB was also operated and characterized in the thermal vacuum environment using measurements taken together with the BCS.

+45˚ Port (AOI=60.5˚)

Space View Port (AOI=11.2˚)

0˚ Port (Nadir)(AOI = 38°)

-45˚ Port (AOI=15.5˚)

BCS

Cryogenic Space View

Source

Scan Mirror

SIS 100

SpMA

Fig. 3. Pre-launch test setup in the thermal vacuum chamber and location of key calibration sources. Pre-launch spectral characterization included measurements of each detector’s relative spectral response (RSR) to in-band (IB) and out-of-band (OOB) radiances, its center wavelength and bandwidth. The IB RSR was made at much smaller wavelength steps than the OOB RSR. The RSB spectral characterization was conducted in ambient while the TEB spectral characterization was performed in the thermal vacuum. A double-grating monochromator, the spectral measurement assembly (SpMA), was used for the MODIS spectral characterization. Special efforts were made to correct for atmospheric transmission, the thermal vacuum (chamber) window’s transmission, and off-axis effects of the SpMA exit slit in order to improve the RSR quality. As shown in Figure 2, the 36 MODIS spectral bands are located on four focal plane assemblies. The band-to-band registration (BBR), detector instantaneous field of view (IFOV), modulation transfer function (MTF), and line spread function (LSF) are the key spatial characterization parameters. They were determined pre-launch using an integration and alignment collimator (IAC). Key pre-launch spectral and spatial references were established and transferred to on-orbit by operating the SRCA together with ground-based equipment [11-15].



Since MODIS collects data over a large scan angle range of ±55º, the system level sensor calibration and characterization activities also included measurements of sensor response versus scan-angle (RVS) for both the TEB and RSB, and scan angle dependent polarization sensitivity for the VIS and NIR spectral bands. When used together, the detector response (1/gain), which is only calibrated at a fixed angle, and the RVS allow data collected at different viewing angles to be fully calibrated. The SIS and BCS were used as the calibration sources for the RSB RVS and TEB RVS, respectively. Polarization sensitivity requirements only apply to the spectral bands with wavelengths below 1µm. The polarization was characterized using a polarization source assembly (PSA). The sensor’s polarization properties are particularly important for the ocean color and key aerosol data products. During RVS and polarization characterization, MODIS was placed on a rotary table with the calibration sources located at fixed positions. More details on the RVS and polarization measurements and results can be found in previous reports [16-18] MODIS RSB on-orbit calibration is reflectance based with reference to its on-board solar diffuser (SD). The SD reference property is its bi-directional reflectance factor (BRF), which was characterized pre-launch by the instrument vendor with traceability to NIST reflectance standards. The BRF characterization was made in a comparison mode using a scattering goniometer at different wavelengths and different viewing geometries. Standard reference samples with NIST traceability were used before and after MODIS SD BRF measurements [19]. Pre-launch calibration and characterization revealed a number of sensor issues, which could impact sensor on-orbit performance. These issues are continuously examined via special calibration efforts, and often require modifications to the general calibration algorithms. For Terra MODIS, the TEB RVS characterization was not successful because of inadequate data sets from pre-launch measurements. Because of this, Terra MODIS TEB RVS had to be characterized post-launch using data collected during spacecraft deep space maneuvers [20]. In addition to identifying noisy and inoperable detectors, pre-launch measurements demonstrated that both Terra and Aqua MODIS SWIR bands (5-7 and 26) were impacted by a thermal leak and electronic crosstalk, although the effect in Aqua MODIS was much smaller than Terra MODIS. An improved and comprehensive on-orbit characterization was necessary to derive and update the crosstalk coefficients and apply corrections in the calibration algorithms. Bands 32-36 of Terra MODIS also experienced a small optical leak from band 31, which also require an on-orbit correction algorithm. Since bands 31-36 consist of photoconductive detectors, this optical leak was referred to as PC crosstalk [21-23]. Two major concerns for Aqua MODIS are the multiple inoperable detectors in band 6 and the BBR for band pairs crossing the cold and warm FPA [24].

4 ON-ORBIT CALIBRATION ACTIVITIES AND METHODOLOGIES On-orbit calibration and characterization activities are routinely performed for both Terra and Aqua MODIS instruments in order to monitor their performance and update calibration look-up tables (LUT) as needed for the level 1B (L1B) data processing. For the RSB, the SD/SDSM observations are made at a frequency varying from weekly to tri-weekly. Typically, each RSB calibration event consists of SD and SDSM measurements over two consecutive orbits, one with and one without the SD screen in front of the SD. Normally, the SD door is closed if no calibration event is scheduled. On Terra MODIS, however, only screened measurements are currently available from a fixed SD door and screen configuration, which resulted from an SD door operation anomaly in May 2003. As Aqua MODIS continues to operate beyond its design life of 6 years, screened measurements are collected every 3 weeks and open measurements every 6 weeks. In support of the RSB calibrations, special spacecraft roll maneuvers are implemented nearly monthly to allow MODIS to view the moon through its SV port. For the TEB, the BB calibration is performed on a scan-by-scan basis. On a



quarterly basis, a BB warm-up/cool-down activity cycles through the BB operational temperature range, providing additional TEB performance information. At mission beginning, the SRCA radiometric mode data was collected monthly, the spectral mode quarterly, and the spatial mode bi-monthly. To preserve lamp usage, spectral and spatial modes are currently performed at reduced frequencies

4.1 RSB calibration The MODIS L1B calibration algorithm converts the instrument scene responses to geolocated and radiometrically calibrated products for all spectral bands. The calibration is band, detector, sub-sample (for 250m and 500m resolution bands), and mirror side dependent. For the RSB, the primary data product is the EV top of the atmosphere (TOA) reflectance factor ρEV·cos(θEV), which is given by [12] ρEV·cos(θEV) = m1·dnEV

*·(dES_EV)2 (1) where θEV is the EV scene solar zenith angle, m1 the calibration coefficient, dES_EV the Earth-Sun distance in AU at the EV observation time, and dnEV

* the EV signal corrected for background, view angle difference, and instrumental temperature effects. From the reflectance factor, the RSB TOA radiance, LEV, can thus be calculated, LEV = eSUN·ρEV ·cos(θEV) /(dES_EV)2 (2)

where eSUN is the solar irradiance normalized by π at the Earth-Sun distance of 1 AU. The solar irradiance for each detector is calculated by integrating the solar irradiance spectrum with its in-band relative spectral response (RSR). The m1 calibration coefficients are derived routinely from SD and SDSM observations. Applying Eqn. 1 to the SD observations, m1 can be expressed by, m1 = ρSD·cos(θSD)/dnSD

*(dES_SD)2 (3)

where ρSD is the SD BRF derived from pre-launch characterization, θSD is the SD solar zenith angle, dnSD

* is the corrected SD detector response, and dES_SD is the Earth-Sun distance in AU at the time of the SD measurement. Practically, two additional factors need to be included to correct for SD on-orbit BRF; degradation and the SD screen (SDS) transmission for the screened measurements [12, 25]. A similar equation can be written for the MODIS lunar observations, which provide independent monitoring of RSB radiometric stability at a different scan angle.

4.2 TEB calibration The L1B primary data product for the TEB is the EV radiance. MODIS TEB are calibrated each scan using a quadratic algorithm, which relates the detector response (dnS) to the sensor at aperture source radiance (LS) after removing the instrument background [13], LS = a0 + b1·dnS +a2·(dnS)2 (4) where a0, b1, and a2 are the offset, linear, and quadratic calibration coefficients. For the EV radiance retrieval, LS includes the EV radiance (LEV) and a scan angle dependent mirror term. Similarly for the BB on-orbit calibration, LS includes the BB source radiance (LBB), a mirror term at a fixed scan angle, and an additional scan cavity term, representing the scan cavity thermal emission reflected from the on-board BB. The radiance terms (BB, scan mirror, and



cavity) are calculated at their corresponding temperatures for each detector using the Planck equation weighted by the RSR. The linear calibration coefficients are computed and updated from the scan-by-scan BB observations at a given BB temperature. The offset and quadratic terms, which have much smaller contributions to the overall calibration algorithm, are derived from quarterly BB warm-up/cool-down activities that provide detector responses over a range of BB temperatures [13, 26].

4.3 Spectral and Spatial Calibration On-orbit spectral calibration measurements can be made for the reflective solar bands with wavelengths less than 1µm when the SRCA is operated in its spectral mode. To the first order, the detector on-orbit relative spectral response, rsr(λ), is proportional to its digital response dn(λ) to the internal SRCA source at different wavelengths λ [14], rsr(λ) = dn(λ)·C(λ) (5) where C(λ) is a wavelength dependent factor computed from the SRCA reference detector’s response during spectral calibration. It is used to correct the SRCA source profile. The lower cased rsr is used here to distinguish it from the pre-launch derived RSR. The critical part of MODIS spectral calibration is to accurately determine the SRCA wavelength. This is accomplished using a didymium filter, which has well characterized spectral transmission peaks. The ratio of SRCA signals before and after the didymium filter is used to determine the source wavelength. Because of the finite width of the SRCA exit slit, the rsr(λ) derived from Eqn. 5 is different from the pre-launch measured RSR(λ) unless an additional correction is applied. Due to SRCA source and signal limitations, on-orbit rsr(λ) profiles only cover the central portion of the pre-launch RSR(λ) profiles. Nevertheless, results derived from the SRCA spectral calibration can be used to track changes in detector center wavelengths and bandwidths over the sensor’s entire mission. Spatial calibrations can also be made by the SRCA when it is configured into its spatial mode. Two reticles of different shapes, one designed for along-scan and another for along-track, are used in the spatial calibration. For each band/detector, a response profile dn(p) as a function of data sample position (p) is produced as the reticle (image) is scanned across the FPA. The centroid of this detector’s response profile or the detector’s position can be determined by [15], <p> = ∑{dn(p)·p}/∑{dn(p)} (6) where the summation is made over all data samples in either along-scan or along-track. The number of effective data samples in the along-scan spatial calibration can be increased by using electronic phase delays of a fraction of the IFOV data sampling time. The spectral band position in the along-scan direction is the average over all detectors. Due to limited data samples, the along-track profile has to be constructed using response profiles from all detectors in a spectral band. Thus only the band position is determined. The band-to-band registration (BBR) between a pair of bands is their position difference. For the along-scan direction, a nominal offset due to band location on the FPA (Figure 2) needs to be removed.

4.4 Special Considerations As discussed in Section 3, there were a few small performance anomalies in both MODIS instruments. Consequently, special considerations have been made during on-orbit calibration in order to mitigate these deficiencies and to remove or reduce their impact on data quality. For both Terra and Aqua MODIS, a SWIR crosstalk correction is implemented in the L1B



calibration algorithm. In addition, a special calibration activity has been designed and carried out periodically to derive and track the correction coefficients [21]. For Terra MODIS, a PC crosstalk correction is applied to bands 32-36. The correction coefficients are derived and monitored using regularly scheduled lunar observations [22]. It is noted that these corrections are applied to both the EV observations and the observations from the OBC used to derive the nominal calibration coefficients. Due to a low gain setting for band 21, which is designed for fire detection, only the fixed linear calibration coefficients are used. For Aqua MODIS bands 33, 35, and 36, fixed linear calibration coefficients are also used, but only during the BB warm-up/cool-down when the BB temperature is above their corresponding saturation temperatures [27].

5 ON-ORBIT PERFORMANCE

5.1 Instrument Operational Activities Major operational activities, significant events and anomalies of Terra MODIS are summarized in Table 2. The Terra spacecraft was launched on December 18, 1999 and the first EV image from MODIS was captured on February 24, 2000 when its nadir aperture door (NAD) opened the first time on-orbit after completing a series of spacecraft and instrument functional check-ups. Since launch, Terra MODIS has operated using four different configurations. As highlighted in Table 2, the instrument initially operated using the A-side electronics and formatter. It was then switched to the B-side electronics and formatter on October 30, 2000. A B-side power supply failure forced the instrument to switch back to the A-side configuration on July, 2, 2001. Shortly after, the A-side formatter failed to operate normally. As a result, a configuration using the A-side electronics cross-strapped with the B-side formatter was used for instrument operations starting on September 17, 2002. Several other notable events occurred for Terra MODIS over its lifetime. Early in the mission the cold FPA slowly lost its temperature control ability due to decreased cooler margin. The temperature control ability was eventually restored after an outgas procedure was performed. Another major event for Terra MODIS was an SD door anomaly in the middle of 2003, which resulted in the SD door being fixed in the open position with its attenuation screen permanently in place. The Aqua spacecraft was launched on May 04, 2002 and the first MODIS image was acquired on June 24, 2002. Unlike Terra MODIS, Aqua MODIS has successfully operated in a single configuration (B-side electronics and formatter) over its entire mission. Table 3 lists key operational activities and significant events for Aqua MODIS. At the mission beginning, Aqua MODIS experienced several safe modes, primarily due to spacecraft related events, but has otherwise operated continuously. Currently, two of the four SRCA 10W lamps have been taken out of operation for both instruments. This has no direct impact on sensor radiometric calibration.

5.2 Instrument Temperature Trends The daily average temperature trends of the instrument, cavity and scan mirror for Terra MODIS are shown in Figure 4a over its on-orbit operation lifetime. As discussed in Section 4, the RSB calibration needs correction for instrument temperature effects and the TEB calibration must consider contributions due to scan mirror and cavity thermal emission. Warm FPA and cold FPA temperature trends are also illustrated in Figure 4b and 4c, respectively. Seasonal oscillations of up to 2K (peak-to-peak) are seen in Figures 4a and 4b. For Terra MODIS, the instrument temperature has increased nearly 3K over its 9 years of on-orbit operation. The cold FPA are well-controlled at a nominal temperature of 83K, with an exception early in the mission when the CFPA lost temperature control. The vertical dashed



lines in the figures indicate major configuration changes and instrument safe mode occurrences. As expected, there are small temperature trend discontinuities between the different configurations. The same temperature trends for Aqua MODIS are shown in Figure 5a, 5b, and 5c. The long-term increase in instrument temperature is approximately 1.5K over the 6.5 years since launch. Overall Aqua cold FPA temperature trends are relatively stable, but show a slight increase of up to 0.1K in recent years due to a gradual decrease of its cooler margin. For nominal operation, the BB is controlled at 290K for Terra MODIS and 285K for Aqua MODIS. On-orbit results show that the BB temperatures have been very stable for both instruments. The long-term drift has been less than 15mK (Terra) and 5mK (Aqua) over the instrument lifetime [28].

Table 2. Significant spacecraft and instrument events for Terra MODIS.

Date Event Description 12/18/99 Terra launch 02/13/00 Science Mode, A-side electronics 02/24/00 First light; nadir aperture door open 06/08/00 Cold focal plane assembly stopped controlling temperature 08/03/00 Focal plane assembly temperature set to 85K 08/05/00 Formatter reset anomaly; MODIS enters standby mode, then safe mode 10/30/00 MODIS switches to B-side electronics configuration 06/15/01 Power supply 2 (PS2) B-side shutdown (Safe Mode) 07/02/01 MODIS switches to A-side electronics configuration using PS1 03/19/02 Spacecraft safe mode hold anomaly during maneuver 09/17/02 Switch to B-side formatter; other components remain on A-side 05/06/03 Solar diffuser door fails to open when commanded 07/02/03 Solar diffuser door set to remain open with screen down 12/16/03 Attitude Control Electronics anomaly; S/C transition to safe mode 11/22/04 SRCA 10W lamp #2 fails to operate normally 02/18/06 SRCA 10W lamp #3 becomes abnormal and is taken out of service 08/22/06 Nadir aperture door and SV door inadvertently closed

Table 3. Significant spacecraft and instrument events for Aqua MODIS.

Date Event Description

05/04/02 Aqua launch 06/07/02 Science Mode, B-side electronics 06/24/02 First light; nadir aperture door open 06/27/02 Spacecraft in safe mode due to a single event upset (SEU) 07/29/02 Spacecraft in safe hold due to spacecraft anomaly 08/09/02 SD door open as command was accidentally dropped; door closed 08/14/02 09/12/02 S/C in safe hold due to ephemeris error; recovered fine pointing same day 04/14/03 SRCA 10W lamp #2 is taken out of service 06/28/05 SRCA 10W lamp #3 is taken out of service



Fig. 4. Daily averaged Terra MODIS telemetry for the instrument, cavity and scan mirror temperatures (a), VIS and NIR FPA (b), and SMIR and LWIR FPA (c). Days with known calibration activities have been removed.

5.3 Noise Characterization MODIS has a total of 490 individual detectors. There were 30 noisy detectors (20 in band 7 and 10 in band 36) identified from pre-launch measurements for Terra MODIS. A noisy detector history of Terra MODIS is provided in Table 4. After more than 9 years of on-orbit operation, there have been 14 new noisy detectors. To some extent, this is expected as the instrument continues aging. It is noted that most of the noisy detectors that became noisy on-orbit are those in the LWIR FPA. Their occurrences are correlated with spacecraft overpasses over the SAA (South Atlantic Anomaly) region. There are no inoperable detectors on Terra MODIS. For Aqua MODIS, pre-launch characterization found that there were only 2 noisy

(a)

(b)

(c)



detectors, but there were 10 inoperable detectors with 8 in band 6. Table 5 is the Aqua MODIS noisy and inoperable detector history. The number of noisy detectors in Aqua MODIS is significantly less than in Terra MODIS. However, multiple inoperable detectors in band 6 are a major concern for a few data products. Because of this, band 6 observations were replaced by band 7 in some applications. Since launch, there have been 5 new noisy detectors and 5 new inoperable detectors (all new inoperable detectors are in band 6).

Fig. 5. Daily averaged Aqua MODIS telemetry for the instrument, cavity and scan mirror temperatures (a), VIS and NIR FPA (b), and SMIR and LWIR FPA (c). Days with known calibration activities have been removed. Currently, most detectors in both instruments continue to meet their design specifications. Their performance, expressed by the band averaged ratios of measured SNR to the specified SNR (RSB) or the specified NEdT to the measured NEdT (TEB), is summarized in Figures 6 and 7. If the ratio is larger than 1, then the detector noise performance meets the specification. Noisy detectors are excluded from the band average unless all detectors in a spectral band are

(a)

(b)

(c)



noisy, such as bands 7 and 36 of Terra MODIS. In addition to those listed in Table 4, Terra MODIS band 8 SNR are currently below the specification as shown in Figure 6. This is largely caused by the decrease of scan mirror reflectance and thus detector response at the typical scene radiance levels. However, the detector intrinsic noise characteristics have changed very little based on the SV observations.

Table 4. Terra MODIS noisy detectors.

Time Event Noisy Band (Detector)

Pre-launch B7(all), B36(all)

2000055.1527 Nadir Door Open B5(4,16), B7(all), B33(1), B34(7,8), B36(all)

2000160.0000 CFPA Lost Control B5(4,16), B7(all), B30(5) B33(1), B34(7,8), B36(all)

2000218.2210 Formatter Anomaly B5(4,16), B7(all), B27(6), B30(5), B33(1), B34(6,7,8), B36(all)

2000304.1420 Switch to B-Side B5(4,16), B7(all), B27(6), B30(5), B33(1), B34(6,7,8), B36(all)

2001019.1415 N/A B5(4,16), B7(all), B27(6), B30(5, 8), B33(1), B34(6,7,8), B36(all)

2001183.2245 Switch to A-Side B5(4), B7(all), B27(6), B30(5, 8), B33(1), B34(6,7,8), B36(all)

2002078.1615 Safe Mode B5(4), B7(all), B27(6), B28(3), B30(5,8), B33(1), B34(5,6,7,8), B36(all)

2003350.1305 Safe Mode B5(4), B7(all), B27(1,6), B28(8), B30(5,8), B33(1), B34(6,7,8), B36(all)

2005130.1345 SAA (Day) B5(4), B7(all), B27(1,6), B28(1,8), B29(6), B30(5,8), B33(1), B34(6,7,8), B36(all)

2005309.1510 N/A B5(4), B7(all), B27(1,6), B28(8,9), B29(6), B30(5,8), B33(1), B34(6,7,8), B36(all)

2006155.0210 SAA (Night) B5(4), B7(all), B27(1,6), B28(8), B29(6), B30(3,5,8), B33(1), B34(6,7,8), B36(all)

2007193.1155 SAA (Day) B5(4), B7(all), B27(1,6), B28(8), B29(6), B30(3,5,8), B33(1), B34(6,7,8), B36(all)

2008308.0900 SAA (Night) B5(4), B7(all), B27(1,2,6), B28(8), B29(6), B30(1,3,5,8), B33(1), B34(6,7,8), B36(all)

Table 5. Aqua MODIS noisy and inoperable detectors.

Time Event Noisy Band (Detector) Inoperable Band (Detector)

Pre-launch B6(17), B20(10) B5(20), B6(2,12-14,16,18-20), B36(5)

2002175.2324 Nadir Door Open B6(7,9,17) B5(20), B6(2,4-6,10,12-16,18-20), B36(5)

2005010.1715 SAA (Day) B6(7,9,17), B27(3) B5(20), B6(2,4-6,10,12-16,18-20), B36(5)

2007359.1020 N/A B6(7,9,17), B27(3), B29(8)

B5(20), B6(2,4-6,10,12-16,18-20), B36(5)

2008038.1750 SAA (Day) B6(7,9,17), B27(3), B29(2,8)

B5(20), B6(2,4-6,10,12-16,18-20), B36(5)



Fig. 6. Terra MODIS pre-launch and on-orbit noise (band averaged).

Fig. 7. Aqua MODIS pre-launch and on-orbit noise (band averaged).

5.4 Radiometric Response The TEB calibration is performed each scan using detector responses to the on-board BB and SV. Their short-term stability can be evaluated through each detector’s scan-by-scan linear response (b1) using a 5-min data granule of over 200 scans. For both Terra and Aqua MODIS, the TEB short-term stability, excluding the noisy detectors identified pre-launch and on-orbit, is generally within ±0.25% over the lifetime of both instruments. The RSB calibration is performed using SD observations when it is illuminated by the Sun. The detector responses used for the short-term stability evaluation must be selected from the SD calibration “sweet” spot, which contains approximately 40 data samples. Excluding effects due to the SD screen, used to calibrate high gain bands, the typical RSB short-term stability, in terms of scan-by-scan calibration coefficient m1, are also within ±0.25% [12, 13]. In addition to an individual detector’s intrinsic characteristics, its response stability could be affected by instrument (or electronics) and FPA temperature stability. Generally, the changing of instrument temperature is a relatively slow process and thus has little impact on short-term stability. If no correction is applied, however, the detector response will show small orbit-to-orbit (daytime and nighttime) and seasonal variations. As illustrated in Figures 4 and 5, the VIS and NIR FPA temperatures are correlated with the instrument temperature. The SMIR and LWIR FPA temperatures are actively controlled. On Terra MODIS, the CFPA temperatures have been very stable. Since mid 2005, Aqua MODIS SMIR and LWIR FPA have gradually seen small orbital temperature variations due to a loss of the cooler’s thermal



control margin. For the most sensitive PC bands (B31-36), the orbital variation in b1 can be up to ±0.5%. Since the TEB calibration is performed on a scan-by-scan basis, the impact due to FPA temperature variations on the resulting L1B product is expected to be minimal. The long-term trends of Terra MODIS radiometric responses, 1/m1 for the RSB and 1/b1 for the TEB, from mirror side 1 are seen in Figure 8 (a) - (e) for the VIS to LWIR spectral bands. Key sensor configuration changes and events are marked via vertical dashed lines. The stepwise changes in Terra MODIS responses are mostly caused by different operational configurations and sensor events. After more than 9 years of on-orbit operation, the VIS bands have experienced large changes in their responses varying from 10-50%. Excluding changes due to different configurations, the response changes are all within 10% for the NIR bands and 5% for the SWIR, MWIR, and LWIR bands. Over a short period when the cooler lost its thermal margin, the CFPA temperatures showed a large increase, which resulted in large changes in the LWIR responses. By comparison, the MWIR bands are less sensitive to their FPA temperature. The long-term radiometric trends from mirror side 2 are very similar to those derived from mirror side 1. There are, however, some differences in terms of the degradation rates in a few VIS spectral bands. Similar long-term radiometric trends for Aqua MODIS are illustrated in Figure 9 (a) - (e). Unlike Terra MODIS, the same configuration has been used in Aqua MODIS operation. Excluding the impact due to a few spacecraft and sensor events that occurred at the beginning of the mission, the Aqua MODIS VIS bands have seen changes of 5-25% in their radiometric responses over a period of more than 6.5 years. The response changes are within 7% for all NIR bands and 2% for all the bands located on the CFPA. In general, the long-term response trending performance of Aqua MODIS has been better (less change) than Terra MODIS. In addition, the mirror side differences in Aqua MODIS are much smaller than Terra MODIS.

5.5 Spectral and Spatial Characterization The Terra and Aqua MODIS spectral characterization performance is summarized in Table 6 for the VIS and NIR spectral bands. Table 6 includes the center wavelengths and bandwidths derived from pre-launch characterization and their on-orbit changes annually averaged at the mission beginning (2000 for Terra MODIS, 2002 for Aqua MODIS) and present (2008). Band 2 results are not provided in Table 6 due to SRCA operation differences between pre-launch and on-orbit spectral measurements. In general, the spectral performance of both Terra and Aqua MODIS has been very stable. Except for Terra MODIS band 8, the changes in the center wavelengths are less than 0.5nm. The changes in bandwidths are also very small, typically less than 0.5nm. Exceptions are bands 1 and 19, which have the largest bandwidths (50nm) among all the VIS and NIR spectral bands. On-orbit spatial performance is examined via band-to-band registration (BBR) in both the along-scan and along-track directions. The 36 spectral bands result in a total of 630 band pairs. The specified spatial registration between any two bands is 200m or less. From pre-launch to on-orbit, nearly all the band pairs in Terra MODIS meet the BBR design requirements. Exceptions are only a few band pairs involving the LWIR bands. For Aqua MODIS, however, serious BBR problems in both along-scan and along-track directions were identified from pre-launch characterization. Though on-orbit performance has been very stable, the BBR problem remains unchanged. If two bands are on the warm FPA (VIS and NIR) or on the cold FPA (SMIR and LWIR), their BBR is likely to meet the specified requirements. Otherwise, if one band is on the warm FPA and the other on the cold FPA, their BBR is typically out of specification. Because of this, special considerations must be made for data products derived from bands in both the warm and cold FPA over non-uniform scenes.



(a)

(b)

(c)

(d)

(e)

Fig. 8. Band averaged gain normalized to initial point for Terra MODIS. Bands are grouped by FPA, a) VIS, b) NIR, c) SWIR, d) MWIR, and e) LWIR. The dashed vertical lines indicate the time of spacecraft and instrument anomalies listed in Table 2.



(a)

(b)

(c)

(d)

(e)

Fig. 9. Band averaged gain normalized to initial point for Aqua MODIS. Bands are grouped by FPA, a) VIS, b) NIR, c) SWIR, d) MWIR, and e) LWIR. The dashed vertical lines indicate the time of spacecraft and instrument anomalies listed in Table 3.



Table 6. MODIS VIS and NIR band center wavelength (CW) and bandwidth (BW) determined from pre-launch characterization and on-orbit changes as measured by the SRCA in 2000 and 2008 (units: nm)

Terra CW and On-orbit Changes Aqua CW and On-orbit Changes

Band CW 2000 2008 CW 2002 2008 1 646.3 0.28 0.31 645.8 0.37 0.49 3 465.7 0.07 0.07 466.1 0.04 0.14 4 553.7 0.05 0.05 553.9 0.03 0.14 8 411.8 -0.59 -0.57 412.4 0.44 0.37 9 442.1 -0.33 -0.24 442.2 0.12 0.23

10 487.0 -0.26 -0.21 487.4 0.09 0.13 11 529.7 -0.11 -0.33 530.1 0.04 0.08 12 546.9 0.05 -0.09 547.2 0.08 0.26 13 665.6 -0.03 -0.41 666.0 0.04 0.18 14 677.0 0.15 0.12 677.6 0.13 0.27 15 746.6 -0.07 -0.30 746.8 -0.14 0.33 16 866.3 -0.18 -0.39 866.9 0.11 0.22 17 904.2 -0.01 0.38 904.4 0.19 0.29 18 935.7 -0.23 -0.47 936.4 0.10 0.28 19 936.2 -0.16 0.09 936.3 0.10 0.74

Terra BW and On-orbit Changes Aqua BW and On-orbit Changes Band BW 2000 2008 BW 2002 2008

1 48.25 0.02 -0.91 47.61 -0.12 0.80 3 18.81 -0.28 0.17 18.87 -0.11 0.17 4 19.81 -0.13 -0.18 19.77 -0.43 -0.01 8 14.80 -0.13 N/A 14.41 -0.09 N/A 9 9.75 -0.37 -0.63 9.67 0.18 0.01

10 10.61 -0.05 -0.15 10.67 -0.01 -0.27 11 12.04 -0.38 -0.54 12.10 0.03 -0.03 12 10.39 -0.34 0.24 10.42 -0.09 -0.24 13 10.10 -0.31 -0.13 10.02 -0.02 0.16 14 11.32 0.42 0.72 11.40 0.21 -0.09 15 9.92 -0.07 0.00 9.79 -0.10 -0.22 16 15.47 0.09 -0.12 15.52 0.26 0.17 17 34.93 -0.20 -0.44 34.94 0.25 -0.17 18 13.60 -0.29 -0.06 13.61 -0.08 -0.27 19 46.13 -0.57 -2.65 46.57 0.90 0.44

5.7 Others MODIS on-orbit lunar observations have been used to track the RSB radiometric calibration stability. Response trends similar to Figures 8 and 9 are produced from the monthly lunar



observations for the VIS and NIR bands [29]. Results from SD and lunar observations are used together to track changes in the sensor response versus scan angle (RVS). For Terra MODIS, the lunar observations are also used to characterize the PC crosstalk, a known problem identified pre-launch. On-orbit changes in SD BRF are tracked regularly by the SDSM [25]. The SD degradation shows distinctive spectral dependence, with the shortest wavelengths having the largest degradation rate. For Terra MODIS, there has been an increase of SD degradation rate at all wavelengths since the middle of 2003. This is due to extra solar exposure as the SD door is kept in its open position. The SD degradation rate of Aqua MODIS is similar to that of Terra MODIS before its SD door anomaly. In addition to using on-board calibrators, the EV targets have also been used to monitor sensor on-orbit performance [30].

6 LESSONS AND APPLICATIONS FOR FUTURE MISSIONS There have been many lessons learned during MODIS design and development, from their pre-launch measurements, and from on-orbit operation and calibration activities [23, 31-32]. A few examples are described in the following, which could provide a useful reference for future missions in terms of instrument design considerations, and pre-launch and on-orbit calibration and validation planning. On the program level, it is extremely important to have a science team formed at the early phase of instrument design and retained throughout the mission. This was particularly the case for the MODIS instruments, as their data are needed by scientists and users from different disciplines. To a large extent, the success of the MODIS program is attributed to the early involvement by the science team (ST), with members selected from different research and application areas including land, ocean, atmosphere, and calibration. After launch, the quality of the data products and the scientific value of the mission must be validated and justified by the combined efforts from the ST members and the broad user community [33-35]. In addition, it is essential to have a dedicated and sustained calibration team working through the entire process, from pre-launch measurements to post-launch operation and calibration. In this case, the MODIS Characterization Support Team (MCST) has provided critical technical support to the Instrument Manager and Project Scientist during sensor pre-launch calibration phases with independent data analysis and assessment of sensor performance. In support of the ST, MCST is also responsible for instrument on-orbit operation and calibration. From a sensor performance perspective, high quality pre-launch measurements of sensor parameters, especially those that cannot be characterized on-orbit, are critical to science data products over the entire mission. The sensor RSR and polarization sensitivity are two examples, which directly impact calibration and retrieval accuracy, yet are extremely difficult to characterize accurately on-orbit. During sensor integration and test (I&T), a few, if not all, unexpected or undesirable features could be exposed and identified with well designed measurements and comprehensive data analyses. The best approach is to eliminate these anomalies with design modifications. If, however, these modifications cannot be made due to cost or schedule constraints, then an attempt must be made to reduce the undesirable effects by exercising and examining different operating conditions, characterizing their impacts, and implementing effective correction algorithms. Aqua MODIS PC optical leak removal, based on lessons from Terra MODIS, and its SWIR crosstalk reduction effort are examples of this technique. It is well known that comprehensive sensor pre-launch calibration and characterization measurements significantly contribute to mission success. Over the years, the MODIS instrument design, including its on-board calibrators, pre-launch calibration and characterization approaches, and on-orbit calibration techniques and algorithms have been incorporated, to some extent, by a number of new earth observing missions. NPOESS VIIRS, GOES-R ABI, and LDCM OLI are just a few examples.



7 SUMMARY This paper has presented an overview of the MODIS instrument background, calibration methodology, on-orbit performance, and lessons learned. MODIS was developed to extend the data records of heritage Earth observing instruments used to study global environmental changes and has produced an unprecedented amount of data for the science community. On-orbit results demonstrate that both MODIS instruments and their OBC are healthy and continue to operate with all designed functions. The temperature control of the BB and cold FPA has been stable and maintained at their nominal operational temperatures for both instruments. In terms of detector performance, of the 490 detectors in each sensor, Terra MODIS currently has no inoperable detectors and a total of 44 detectors that are below the SNR requirement (including all detectors for bands 7 and 36 identified pre-launch), and Aqua MODIS has 15 inoperable detectors (13 in band 6 identified pre-launch and at mission beginning) and 6 detectors that fail to meet the SNR requirement. Overall, the long-term radiometric response for both MODIS instruments has been well characterized and corrected in the L1B algorithms. On Terra, the largest response changes are seen in the VIS bands, up to nearly 50% for band 8 (shortest wavelength). The Terra NIR bands response changes are less than 10% and the SMIR and LWIR bands less than 5%. Aqua MODIS radiometric response changes are smaller than those for Terra MODIS, 5-25% for the VIS bands, up to 7% for the NIR bands and less than 2% for the SMIR and LWIR bands. The spectral and spatial performance of both MODIS instruments has been stable since launch. Overall, the MODIS instruments continue to produce high quality science data. The continuing efforts of MCST are essential as both instruments continue to operate beyond their design life. Lessons learned and experience gained from MODIS in all phases - design, development, pre-launch and on-orbit - are a useful reference for future missions.

APPENDIX A: LIST OF ACRONYMS AIRS – Atmospheric Infrared Sounder AMSR-E – Advanced Microwave Scanning Radiometer for EOS AMSU – Advanced Microwave Sounding Unit AOI – Angle of Incidence ASTER – Advanced Spaceborne Thermal Emission and Reflection Radiometer AVHRR – Advanced Very High Resolution Radiometer BB – Blackbody BBR – Band-to-band Ratio BCS – Blackbody Calibration Source BOL – Beginning of Life BRF – Bi-directional Reflectance Factor CALIPSO – Cloud-Aerosol Lidar and Infrared Pathfinder Satellite Observation CERES – Clouds and the Earth’s Radiant Energy System CZCS – Coastal Zone Color Scanner EO-1 – Earth Observing Mission 1 EOL – End of Life EOS – Earth Observing System EV – Earth View FM1 – Flight Model 1 FPA – Focal Plane Assembly GOES-R ABI – Geostationary Operational Environmental Satellite R-series Advanced

Baseline Imager GSFC – Goddard Space Flight Center HIRS – High Resolution Infrared Radiation Sounder HSB – Humidity Sounder for Brazil



I&T – Integration and Test IAC – Integration and Alignment Collimator IB – In-Band IFOV – Instantaneous Field of View Landsat TM – Landsat Thematic Mapper LDCM OLI – Landsat Data Continuity Mission Operational Line Imager LSF – Line Spread Function LST – Local Standard Time LUT – Look-Up Table LWIR – Long-wave Infrared (Bands 27-36) MCST – MODIS Characterization Support Team MISR – Multi-angle Imaging Spectro-Radiometer MODIS – Moderate Resolution Imaging Spectroradiometer MOPITT – Measurements of Pollution in the Troposphere MTF – Modulation Transfer Function MWIR – Middle-wave Infrared (Bands 20-25) NIR – Near Infrared (Bands 1, 2, 13-19) NPOESS VIIRS – National Polar-orbiting Operational Environmental Satellite System

Visible Infrared Imager Radiometer Suite OBC – On-Board Calibrator OOB – Out-of-Band PARASOL – Polarization & Anisotrophy of Reflectances for Atmospheric Sciences coupled

with Observations from a Lidar PC – Photoconductive PFM – Protoflight Model PSA – Polarization Source Assembly PV – Photovoltaic RSB – Reflective Solar Bands RSR – Relative Spectral Response RVS – Reflectivity Versus Scan Angle SBRS – Santa Barbara Remote Sensing SD – Solar Diffuser SDS – Solar Diffuser Screen SDSM – Solar Diffuser Stability Monitor SIS – Spectral Integration Sphere SMIR – Short- and Mid-Wave Infrared (Bands 5-7, 20-26) SNR – Signal to Noise Ratio SpMA – Spectral Measurement Assembly SRCA – Spectral Radiometric Calibration Assembly ST –Science Team SV – Space View SWIR – Short-wave Infrared (Bands 5-7, 26) TDI – Time Delay and Integration TEB – Thermal Emissive Bands TOA – Top of Atmosphere VIS – Visible (Bands 3, 4, 8-12)

Acknowledgments The authors thank the MODIS Characterization Support Team, in particular Amit Angal, Taeyoung Choi, and Na Chen, for their contributions and data analysis.



References [1] W.L. Barnes and V.V. Salomonson, "MODIS: A global image spectroradiometer for

the Earth Observing System," Critical Rev. Optical Sci. Technol. CR47, 285-307 (1993).

[2] W. L. Barnes, X. Xiong, B. W. Guenther, and V. Salomonson, "Development, characterization, and performance of the EOS MODIS sensors," Proc. SPIE 5151, 337-345 (2003) [doi: 10.1117/12.504818].

[3] V. V. Salomonson, W. Barnes, and E. J. Masuoka, "Introduction to MODIS and an overview of associated activities," in Earth Science Satellite Remote Sensing, Vol.1: Science and Instruments, J. J. Qu, Ed., pp. 12-32, Springer, New York (2006).

[4] V. V. Salomonson, W. L. Barnes, X. Xiong, S. Kempler, and E. Masuoka, "An overview of the Earth Observing System MODIS instrument and associated data systems performance," Int. Geosci. Rem. Sens. Symp. 2, 1174-1176(2002) [doi:10.1109/IGARSS.2002.01025812].

[5] C. L. Parkinson, "Aqua: An earth-observing satellite mission to examine water and other climate variables," IEEE Trans. Geosci. Rem. Sens. 41, 173-183 (2003) [doi:10.1109/TGRS.2002.808319].

[6] C. O. Justice, E. Vermote, J. R. G. Townshend, R. Defries, D. P. Roy, D. K. Hall, V. V. Salomonson, J. L. Privette, G. Riggs, A. Strahler, W. Lucht, R. B. Myneni, P. Lewis, and M. J. Barnsley, "The moderate resolution imaging spectroradiometer (MODIS): land remote sensing for global change research," IEEE Trans. Geosci. Rem. Sens. 36, 1228-1249 (1998) [doi:10.1109/36.701075].

[7] W. E. Esaias, M. R. Abbott, I. Barton, O. W. Brown, J. W. Campbell, K. L. Carder, D. K. Clark, R. L. Evans, F. E. Hoge, H. R. Gordon, W. P. Balch, R. Letelier, and P. J. Minnett, "An overview of MODIS capabilities for ocean science observations," IEEE Trans. Geosci. Rem. Sens. 36, 1250-1265 (1998) [doi:10.1109/36.701076].

[8] M. D. King, W. P. Menzel, Y. J. Kaufman, D. Tanre, B. C. Gao, S. Platnick, S. A. Ackerman, L. A. Remer, R. Pincus, and P. A. Hubanks, "Cloud and aerosol properties, precipitable water, and profiles of temperature and water vapor from MODIS," IEEE Trans. Geosci. Rem. Sens. 41, 442-458 (2003) [doi:10.1109/TGRS.2002.808226].

[9] X. Xiong, K. Chiang, J. Esposito, B. Guenther, and W. Barnes, "MODIS on-orbit calibration and characterization," Metrologia 40, 89-92 (2003) [doi:10.1088/0026-1394/40/1/320].

[10] X. Xiong and W. L. Barnes, "An overview of MODIS radiometric calibration and characterization," Adv. Atmos. Sci. 23(1), 69-79 (2006) [doi:10.1007/s00376-006-0008-3].

[11] W. L. Barnes, T. S. Pagano, and V. V. Salomonson, "Prelaunch characteristics of the moderate resolution imaging spectroradiometer (MODIS) on EOS-AM1,"IEEE Trans. Geosci. Rem. Sens. 36, 1088-1100 (1998) [doi:10.1109/36.700993].

[12] X. Xiong, J. Sun, W. Barnes, V. Salomonson, J. Esposito, H. Erives, and B. Guether, "Multiyear on-orbit calibration and performance of Terra MODIS reflective solar bands," IEEE Trans. Geosci. Rem. Sens. 45(4), 879-889 (2007) [doi:10.1109/TGRS.2006.890567].

[13] X. Xiong, K. Chiang, A. Wu, W. Barnes, B. Guether, and V. Salomonson, "Multiyear on-orbit calibration and performance of Terra MODIS thermal emissive bands," IEEE Trans. Geosci. Remote Sensing, 46 (6), 1790-1803 (2008) [doi:10.1109/TGRS.2008.916217].

[14] X. Xiong, N. Che, and W.L. Barnes, "Terra MODIS on-orbit spectral characterization and performance," IEEE Trans. Geosci. Rem. Sens. 44(8), 2198-2206 (2006) [doi:10.1109/TGRS.2006.872083].



[15] X. Xiong, N. Che, and W.L. Barnes, "Terra MODIS on-orbit spatial characterization and performance," IEEE Trans. Geosci. Rem. Sens. 43, 355-365 (2005) [doi:10.1109/TGRS.2004.840643].

[16] C. Pan, X. Xiong, and N.Che, "MODIS pre-launch characterization of reflective solar band response vs. scan angle," Proc. SPIE 6677, 66770R (2007) [doi:10.1117/12.730573].

[17] K. Chiang and X. Xiong, "Pre-launch characterization of Aqua MODIS scan mirror response versus scan angle for thermal emissive bands," Proc. SPIE 6677, 66771J (2007) [doi:10.1117/12.731562].

[18] J. Sun and X. Xiong, "MODIS polarization sensitivity analysis," IEEE Trans. Geosci. Rem. Sens. 45 (9), 2875-2885 (2007) [doi:10.1109/TGRS.2007.900672].

[19] X. Xiong, J. Sun, J. Esposito, X. Liu, W. L. Barnes, and B. Guenther, "On-orbit characterization of a solar diffuser’s bidirectional reflectance factor using spacecraft maneuvers," Proc. SPIE 5151, 375-383 (2003) [doi:10.1117/12.504802].

[20] X. Xiong, V. V. Salomonson, K. Chiang, A. Wu, B. W. Guenther, and W. Barnes, "On-orbit characterization of RVS for MODIS thermal emissive bands," Proc SPIE 5652, 210 (2004) [doi:10.1117/12.578344].

[21] X. Xiong, K. Chiang, F. Adimi, W. Li, H. Yatagai, and W. L. Barnes, "MODIS correction algorithm for out-of-band response in the short-wave IR bands," Proc. SPIE 5234, 605-613, (2003) [doi:10.1117/12.510261].

[22] W. Li, X. Xiong, K. Chiang, and G. Toller, "Evaluation of Terra MODIS PC bands optical leak correction algorithm," Proc. SPIE 5882, 588219 (2005) [doi:10.1117/12.614509].

[23] X. Xiong, W. L. Barnes, B. Guenther, and R.E. Murphy, "Lessons learned from MODIS," Adv. Space Res. 32(11), 2107-2112 (2003)[doi:10.1016/S0273-1177(03)90530-8].

[24] X. Xiong, K. Chiang, J. Sun, N. Che, and W. L. Barnes, "Aqua MODIS first year on-orbit calibration and performance," Proc. SPIE 5234, 391-399 (2003) [doi:10.1117/12.510580].

[25] X. Xiong, H. Erives, S. Xiong, X. Xie, J. Esposito, J. Sun, and W. Barnes, "Performance of Terra MODIS solar diffuser and solar diffuser stability monitor," Proc. SPIE 5882, 58820S(2005) [doi:10.1117/12.615334].

[26] X. Xiong, N. Chen, S. Xiong, K. Chiang, and W. Barnes, "Performance of the Terra MODIS on-board blackbody," Proc. SPIE 5882, 58820U(2005) [doi:10.1117/12.615431].

[27] X. Xiong, B. N. Wenny, A. Wu, W. L. Barnes, and V. V. Salomonson, "Aqua MODIS thermal emissive band on-orbit calibration, characterization, and performance," IEEE Trans. Geosci. Rem. Sens. 47(3), 803-814 (2009) [doi:10.1109/TGRS.2008.2005109].

[28] X. Xiong, B. N. Wenny, A. Wu, and W. Barnes, "MODIS On-board blackbody function and performance," IEEE Trans. Geosci. Rem. Sens. (in review), (2009).

[29] J. Sun, X. Xiong, W. L. Barnes, and B. Guenther, "MODIS reflective solar bands on-orbit lunar calibration," IEEE Trans. Geosci. Rem. Sens. 45(7), 2383-2393(2007) [doi:10.1109/TGRS.2007.896541].

[30] X. Xiong, A. Wu, and B. N. Wenny, "Using dome C for MODIS calibration stability and consistency," J. Appl. Rem. Sens. 3, 033520, (2009) [doi:10.1117/1.3116663].

[31] X. Xiong, N. Che, C. Pan, X. Xie, J. Sun, W. L. Barnes, and B. Guenther, "Results and lessons form MODIS reflective solar bands calibration: pre-launch to on-orbit," Proc. SPIE 6296, 629607 (2006) [doi:10.1117/12.679144].

[32] X. Xiong, K. Chiang, N. Chen, S. Xiong, W.L. Barnes, and B. Guenther, "Results and lessons from MODIS thermal emissive bands calibration: pre-launch to on-orbit," Proc. SPIE 6296, 62960A (2006) [doi:10.1117/12.680992].



http://dx.doi.org/10.1016/S0273-1177(03)90530-8

http://dx.doi.org/10.1016/S0273-1177(03)90530-8

[33] Z. Wan, Y. Zhang, Q. Zhang, and Z.-L. Li, "Quality assessment and validation of the MODIS Global Land Surface Temperature," Int. J. Rem. Sens. 25(1), 261-278 (2004) [doi:10.1080/0143116031000116417].

[34] P. J. Minnett, R.H. Evans, E. J. Kearns, and O. B. Brown, "Sea-surface temperature measured by the moderate resolution imaging spectrometer (MODIS)," IEEE Int. Geosci. Rem. Sens. Symp. 2, 1177-1179 (2002) [doi:10.1109/IGARSS.2002.1025872].

[35] S. Hook, R. Vaughan, H. Tonooka, and S. Schladow, "Absolute Radiometric In-Flight Validation of Mid and Thermal Infrared Data from ASTER and MODIS Using the Lake Tahoe CA/NV, USA Automated Validation Site," IEEE Trans. Geosci. Rem. Sens. 45(6), 1798-1807 (2007) [doi:10.1109/TGRS.2007.894564].



http://dx.doi.org/10.1109/IGARSS.2002.1025872

Overview of NASA Earth Observing Systems Terra and Aqua ...€¦ · [email protected] cJoint...

Documents

Transcript of Overview of NASA Earth Observing Systems Terra and Aqua ...€¦ · [email protected] cJoint...