Minimize OCO-2

OCO-2 (Orbiting Carbon Observatory-2)

OCO-2 is a NASA mission dedicated to studying atmospheric carbon dioxide. Carbon dioxide is the leading human-produced greenhouse gas driving changes in the Earth's climate. OCO-2 will provide a complete picture of human and natural carbon dioxide sources and "sinks," the places where the gas is pulled out of the atmosphere and stored. The aim is to map the global geographic distribution of these sources and sinks and study their changes over time. 1) 2) 3) 4) 5)

The OCO-2 mission is expected to provide answers to such questions as: “What controls atmospheric carbon dioxide (CO2)” ?

• Natural systems including the ocean and plants on land both absorb and emit carbon dioxide to the atmosphere

• Currently, these natural systems are absorbing about half of the carbon dioxide emitted by human activities

• These natural carbon dioxide “sinks” are limiting the rate of carbon dioxide buildup and its impact on the Earth’s climate

• So far it is not known:

- Exactly where the carbon dioxide is being emitted and absorbed

- How much longer natural processes will continue to absorb the carbon dioxide that we emit in the presence of climate change.

OCO-2 is a follow-up spacecraft to OCO-1, lost during a launch vehicle failure on Feb. 24, 2009. In the days following the devastating loss of the OCO-1 mission, the JPL project team reassembled to respond to requests from NASA/HQs about the state of spares that could be used for an OCO re-flight, and the costs and schedule for building another OCO observatory. These initial studies included options to launch OCO-2 on a different platform, including the options of sharing the platform with other instruments, co-manifested launches and single launch missions.

The loss of the OCO mission as a result of a launch vehicle payload fairing anomaly resulted in a setback for the international carbon cycle science community. There have been and continue to be advances made leading to improved understanding of the global climate change process. However, the current set of ground-based, airborne, and spaceborne instrument and sensor measurements do not allow carbon sources and sinks to be quantified and examined over time. The OCO instrument is the only confirmed sensor that addresses not only science, but policy imperatives as well. A four month-long study to examine options for recovery of the critical scientific measurement resulted in a decision to rebuild the original OCO observatory to the extent possible.

In Feb. 2010, a budget could be secured providing funding to support a launch of the OCO re-flight mission (now known as OCO-2) no later than February 2013. The OCO-2 project is managed by NASA/JPL (Jet Propulsion Laboratory) in Pasadena, CA.

Justification for an OCO re-flight mission:

• Accurate and precise measurements of carbon dioxide sources and sinks are of paramount importance

• Despite progress, our knowledge is limited by the lack of high precision global measurements of atmospheric carbon dioxide

• While there have been advances in space-based measurements there is no existing or confirmed sensor capable of quantifying carbon dioxide sources and sinks.

• A re-flight of a “carbon copy” of OCO meets the science and policy imperatives at the lowest cost, and on the fastest possible schedule.

Table 1: OCO-2 is a re-flight of the OCO-1 mission 6)


Figure 1: Artist's rendition of the OCO-2 spacecraft (image credit: OSC)


OCO-2 is based on the Orbiting Carbon Observatory (OCO) to the extent possible. OCO-2 will be a dedicated spacecraft to fly a single instrument comprised of three high resolution grating spectrometers. 7) 8)

The spacecraft is three-axis zero-momentum stabilized, using the LeoStar-2 bus of OSC [LeoStar-2 is of OCO, SORCE (Solar Radiation and Climate Experiment) and GALEX (Galaxy Evolution Explorer) heritage]. The bus houses and points the instrument, provides power, receives and processes commands from the ground, records, and downlinks the data collected by the instrument, and maintains its position with the EOS A-Train. The primary structure consists of a 2.12 m long hexagonal column that is 0.94 m wide.

The ACS (Attitude Control Subsystem) points the instrument for science and calibration observations and the body-mounted X-band antenna at the ground station for data downlink. Pitch, roll, and yaw are controlled by 4 reaction wheels. Three magnetic torque rods are used to de-spin the reaction wheels. OCO-2 uses the same types of Goodrich/Ithaco reaction wheels and torque rods that were used for OCO. However, the reaction wheels have been modified to address lifetime issues identified over the past decade.

The spacecraft orbit is determined by a GPS receiver of General Dynamics Viceroy. Pointing information is provided by a star tracker (Sodern), a MIMU (Miniature Inertial Measurement Unit) of Honeywell, and a magnetometer of Goodrich.

EPS (Electrical Propulsion Subsystem): Two deployable solar panels supply ~900 W when illuminated at near normal incidence (use of GaAs solar cells). The solar panels charge an Eagle Picher 35 Ah NiH (Nickel-Hydrogen) battery that provides power during eclipse. OCO-2 uses the same battery model used by OCO.

The central electronics unit uses a RAD-6000 flight computer of BAE to manage the attitude control, power, propulsion, and telecom systems, and the 128 Gbit solid-state recorder of Seaker that stores the science data collected by the instrument. The RAD6000 had to be modified to replace the SRAM (Static Random Access Memory) part, which was based on an obsolete, 16-chip MCM (Multi-Chip Module) with a newer SRAM part.

A propulsion subsystem will be used for orbit maintenance (use of 45 kg of hydrazine). Once in orbit, it is being used to raise and adjust the orbit altitude and inclination as necessary to maintain the spacecraft’s position in the A-Train. Finally, it is used to de-orbit the Observatory at the end of the mission.

RF communications: Both science and housekeeping data are usually returned to the ground at 150 Mbit/s using an L3 Communications X-band transmitter and a body-mounted X-band patch antenna. Spacecraft and instrument housekeeping data can also be returned by an S-band transmitter to a ground station or through a NASA TDRS (Tracking and Data Relay Satellite) system. The uplink is implemented with an S-band receiver and a pair of omni-directional antennas. OCO-2 will use S-band hardware from TAS (Thales Alenia Space) that meets the same performance requirements as OCO, but is based on a new, all-digital design.

Spacecraft bus

LeoStar-2 bus of OSC (LeoStar-2 is of OCO-1, SORCE and GALEX mission heritage)

Spacecraft launch mass

530 kg

Solar arrays

Triple junction GaAs, NiH battery with a capacity of 35 Ah

Spacecraft power

521 W orbit average

Spacecraft stabilization

3-axis, zero momentum



Mission design life

2 years


Altitude of 705 km, flying in polar, sun-synchronous formation with the EOS A-train

RF communications

S-band transceiver and X-band science data transmitter

Table 2: Overview of OCO-2 minisatellite parameters


Figure 2: Photo of the OCO-2 spacecraft during integration and test at OSC (image credit: OSC, Ref. 7)


• The OCO-2 mission underwent CDR (Critical Design Review) in August 2010 and key design point-C (KDP-C) in September 2010 (Ref. 1).

• The OCO-2 instrument will be delivered in the spring of 2012 (Ref. 17).


Launch: The launch of OCO-2 is scheduled for July 2014 on a Delta-2 vehicle from VAFB, CA. NASA contracted ULA (United Launch Services LLC) in July 2012. 9)

Note: Originally, NASA planned a launch of the OCO-2 mission for Feb. 2013 on a Taurus- XL 3110 vehicle of OSC (However, the configuration Taurus-XL and the OCO-1 spacecraft experienced already a launch failure on Feb. 24, 2009. This was followed by a launch failure of NASA's Glory spacecraft on a Taurus-XL 3110 vehicle on March 4, 2011). — Hence, in early 2012, NASA and OSC mutually agreed to terminate the existing launch contract for OCO-2. In early February 2012, NASA released a multi-mission request for launch service proposals that included the OCO-2 mission. Once NASA officials select a new rocket for OCO-2, it will involve a launch delay at least into the year 2014. 10) 11)

Orbit: Sun-synchronous orbit, altitude = 705 km, inclination = 98.2º, period = 98.8 min, repeat cycle of 16 days, equatorial crossing time at 13:15 hours on an ascending node.

OCO-2 will initially be launched into an orbit of 635 km altitude. The on-board propulsion system will then raise the orbit to 705 km. The OCO-2 spacecraft will become part of the “A-Train” (a loose formation of the Afternoon Constellation consisting of: Aqua, CloudSat, CALIPSO, Glory, PARASOL, Aura, and OCO-2) to correlate the OCO-2 data with data acquired by other instruments such as AIRS on Aqua. OCO-2 will fly ahead of the A-train, about 15 minutes before the Aqua spacecraft.


Figure 3: Artist's view of the OCO-2 spacecraft flying in the A-Train (image credit: NASA/JPL)


Sensor complement: (OCO-2 instrument)

The OCO-2 instrument is almost a copy of the OCO instrument, designed and developed by HSSS (Hamilton Sundstand Sensor Systems), Pomona, CA. The only changes being made to the instrument are associated with parts obsolescence or mitigation of known performance issues (Table 3). 12) 13) 14) 15)

The objective of the OCO-2 spectrometers is to measure sunlight reflected off the Earth's surface. The rays of sunlight that enter the spectrometers will pass through the atmosphere twice, once as they travel from the Sun to the Earth, and then again as they travel from the Earth's surface to the OCO-2 instrument. Carbon dioxide and molecular oxygen molecules in the atmosphere absorb light energy at very specific colors or wavelengths. Thus, the light that reaches the OCO-2 instrument will display diminished amounts of energy at those characteristic wavelengths. The OCO-2 instrument employs a diffraction grating to separate the inbound light energy into a spectrum of multiple component colors. The reflection gratings used in the OCO-2 spectrometers consist of a very regularly-spaced series of grooves that lie on a very flat surface. The back of a compact disc is an everyday example of a diffraction grating. 16)

The characteristic spectral pattern for CO2 can alternate from transparent to opaque over very small variations in wavelength. The OCO-2 instrument must be able to detect these dramatic changes, and specify the wavelengths where these variations take place. Thus, the grooves in the instrument diffraction grating will be very finely tuned to spread the light spectrum into a large number of very narrow wavelength bands or colors. Indeed, the OCO-2 instrument design incorporates 17,500 different colors to cover the entire wavelength range that can be seen by the human eye. A digital camera covers the same wavelength range using just three colors.

The OCO-2 experiment requires the measurement of three relatively small bands of electromagnetic radiation, where the spectral wavelength ranges of these three critical bands are widely separated. To accomplish this task economically, OCO-2 employs three spectrometers instead of one. Each spectrometer will measure light in one specific region of the spectrum. The focal plane associated with each spectrometer is designed to detect very fine differences in wavelength within each of these spectral ranges.

OCO-2 measurements must be very accurate. To eliminate energy from other sources that would generate measurement errors, the light detectors for each camera must remain very cold. To ensure that the detectors remain sufficiently cold, the OCO-2 instrument design will include a cryocooler, which is a refrigeration device. The cryocooler keeps the detector temperature at or near -120° C.

Performance obsolescence:

• When work began to evaluate a reflight of the OCO design, there were several electronics parts that were identified to be out of production. Flight worthy residual parts or direct replacements have been identified for all of these.

• Of much larger concern was the cryocooler used to maintain the Focal Plane Arrays at their operating temperatures (120 K). OCO used the flight spare cryocooler original purchased for the TES (Tropospheric Emission Spectrometer) that is currently flying on NASA’s Aura mission. This cooler was purchased in the late 1990’s and was no longer in production. After months of research, the project located a spare two-stage, split tube cryocooler assembly belonging to the NOAA GOES-R (Geostationary Operational Environmental Satellite Series R) program that could be modified to OCO-2 specifications. It was procured for the ABI (Advanced Baseline Imager) instrument. A NASA and NOAA inter-agency exchange MOA (Memorandum of Agreement) allowed the transfer of the GOES-R cryocooler assembly to the OCO-2 project. The OCO-2 project is procuring a replacement unit that will be returned to the GOES-R program when it is complete. While the ABI cryocooler has similar capacity to the TES flight spare used on the original OCO instrument, it is much smaller and has a different set of electronics used to operate it.

• The smaller size required a new bracket to hold the new cooler in the correct position, a slight change to the variable conductance heat pipes used to remove the waste heat and small positional changes in the thermal links between the cryocooler’s coldhead and the focal plane arrays.

• The new electronics presented two challenges. First, the ABI cryocooler communication interface uses low-voltage differential signaling (LVDS) instead of the TES cryocooler’s RS-422. To handle this, a small electronics board, i.e. the CCIE (Cryocooler Interface Electronics), has been developed to buffer communication and retransmit it after converting between the RS-422 / LVDS protocols. The second challenge is that the command and telemetry formats are completely different. The CCIE will recognize the incoming TES cryocooler commands and reformat them and/or substitutes the equivalent commands for the ABI cooler. This approach will avoid the need to either the spacecraft or instrument electronics or flight software – allowing OCO-2 to use the same designs that were qualified for the original mission.

Performance anomalies:

• In addition, there were a handful of performance anomalies that required correction in the ground processing: residual image, slit misalignment and spectral stray light. Removing these anomalies from the instrument will greatly simplify the complexity of the ground data processing and improve the final quality of the science data.

- Residual image: The OCO-2 project has taken two paths to remove the residual image from the next instrument: new detectors and/or a lower operating temperature.

The original OCO instrument used a silicon detector for the O2 A-band: the Teledyne Scientific & Imaging, LLC HyVISI. These had identical readout to the Hawaii-1RG mercury-cadmium-telluride (HgCdTe) detectors used in the CO2 bands – allowing a common design for the OCO electronics to provide the bias, timing and readout. The HgCdTe detectors were not an option for the O2 A-band as the quantum efficiency at 0.76 µm was much too low. To save power, the O2 A-band detector was only cooled to 180 K as the silicon dark current is much less than the HgCdTe detectors.

Early in OCO-2 planning, the option of using substrate removed HgCdTe was explored to provide good sensitivity at 0.76 µm. One lot of parts was procured and initial testing looks likely that these parts will meet the OCO-2 needs.

- Slit misalignment: The OCO slit misalignment is thought to be the result of applying final torque to bolts that control the alignment of the slits to the optical bench. This created a situation where energy was stored in the bolts to be released during the vibration testing. For OCO-2, the installation order of these bolts will be corrected to avoid storing any energy. Once the spectrometers’ assemblies are all installed in the optical bench, a seating vibration test will be conducted to validate the slits do not move. This will allow validation of this approach to correct the slit misalignment before the instrument level vibration testing.

- Spectral stray light: It is now believed that the substrate in the original HgCdTe detectors provided the coupling of the stray light into the science area of the focal planes. The focal plane test bed mentioned above has been used to recreate the stray light problem in spare HgCdTe detectors from OCO, while the silicon detectors do not exhibit the problem. When the first substrate-removed HgCdTe detectors are tested (expected in mid-September 2010), it is expected that they will not exhibit the stray light problems.

Table 3: Changes planned for the OCO-2 instrument


Figure 4: Illustration of the OCO-2 instrument (image credit: HSSS/JPL)

As for OCO, the OCO-2 instrument consists of three, co-bore-sited, long-slit, imaging grating spectrometers optimized for the O2 A-band at 0.765 µm and the CO2 bands at 1.61 and 2.06 µm (Figure 6). The 3 spectrometers use similar optical designs and are integrated into a common structure to improve system rigidity and thermal stability. They share a common housing and a common Cassegrain telescope (Ref. 3).

The telescope consists of an 11 cm aperture, as well as a primary and a secondary mirror. The relay optics assembly includes fold mirrors, dichroic beam splitters, band isolation filters and re-imaging mirrors. Each spectrometer consists of a slit, a two-lens collimator, a grating, and a two-lens camera. Each of the three spectrometers has essentially an identical layout. Minor differences among the spectrometers, such as the coatings, the lenses and the gratings, account for the different bandpasses that are characteristic of each channel. The focal ratios of the instrument optics will range from f/1.6 to f/1.9.

To implement an optically fast, high-spectral-resolution measurement system, the OCO-2 instrument combines refractive and reflective optical techniques. Since the light in the common telescope and relay optics assembly will not separate into the three distinct wavelength bands, these instrument subsystems use primarily reflective optics. On the other hand, the extremely narrow channel bandpasses make potential chromatic aberrations in the spectrometers negligible, which enable the use of refractive optics.

The light path is illustrated in Figures 5 and 6. Light entering the telescope is focused at a field stop and then re-collimated before entering a relay optics assembly. There, it is directed to one of the three spectrometers by a dichroic beam splitter, and then transmitted through a narrowband pre-disperser filter. The pre-disperser filter for each spectral range transmits light with wavelengths within ~±1% of the central wavelength of the CO2 or O2 band of interest and rejects the rest. The light is then refocused on the spectrometer slits by a reverse Newtonian telescope.


Figure 5: Illustration of OCO grating spectrometers and scheme of optical system (image credit: JPL/HSSS)

Each spectrometer slit is about 3 mm long and about 25 µm wide. These long, narrow slits are aligned to produce co-bore-sited fields of view that are ~0.0001 radians wide by ~0.0146 radians long. Because the diffraction gratings efficiently disperse only the light that is polarized in the direction parallel to the slit, a polarizer was included in front of the slit to reject the unwanted polarization before it enters the spectrometer, where it could contribute to the scattered light background.


Figure 6: The OCO-2 instrument showing the major optical components and optical path (image credit: NASA/JPL) 17)

Once the light traverses a spectrometer slit, it is collimated by a 2-element refractive collimator, dispersed by a gold-coated, reflective planar holographic diffraction grating, and then focused by a 2-element camera lens on a 2-dimensional FPA (Focal Plane Array), after traversing a second, narrowband filter. The narrowband filter just above the FPA is cooled to ~180 K to reject thermal emission from the instrument.

Following the OCO design, the spectral range and resolving power of each channel includes the complete molecular absorption band as well as some nearby continuum to provide constraints on the optical properties of the surface and aerosols as well as absorbing gases. To meet these requirements, the O2 A-band channel covers 0.758 to 0.772 µm with a resolving power of > 17,000, while the 1.61 and 2.06 µm CO2 channel cover 1.594 to 1.619 µm and 2.042 to 2.082 µm, respectively with a resolving power > 20,000.

The spectrometer optics project a 2-dimensional image of a spectrum on 1024 by 1024 pixel FPA with 18 µm pixels (Figure 7). The grating disperses the 1024-pixel wide spectrum in the direction perpendicular to the long axis of the slit. The full-width at half maximum (FWHM) of the slit image on the FPA is sampled by 2 to 3 pixels in the direction of dispersion. The length of the slit limits spatial field of view to only ~190 pixels in the dimension orthogonal to the direction of dispersion. Science measurements are restricted to the center ~160 of these 190 pixels.

For normal science operations, the FPAs are continuously read out at 3 Hz. To reduce the downlink data rate and increase the signal to noise ratio, ~20 adjacent pixels in the FPA dimension parallel to the slit (i.e. the “spatial direction” in Figure 7) are summed on board to produce up to 8 spatially-averaged spectra along the slit. The along-slit angular field of view of each of these spatially-averaged “super-pixels: is ~1.8 mrad (0.1º or ~1.3 km at nadir from a 705 km orbit). The angular width of the narrow dimension of the slit is only 0.14 mrad, but the focus of the entrance telescope was purposely blurred to increase the effective full width at half maximum of each slit to ~0.6 mrad to simplify the boresight alignment among the 3 spectrometer slits.


Figure 7: The illumination and readout scheme used for the OCO-2 FPAs (image credit: NASA/JPL)

In addition to the 8 spatially-binned, 1024-element spectra, each spectrometer also returns 4 to 20 spectral samples without on-board spatial binning to provide the full along-slit spatial resolution. Each of these full-resolution “color stripes” covers a 220 pixel wide region of the FPA that includes the full length of the slit (190 pixels) as well as a few pixels beyond the ends of the slit (Figure 7). These full-spatial-resolution color stripes are used to detect spatial variability within each of the spatially summed super pixels and to quantify the thermal emission and scattered light within the instrument.

For the OCO instrument, the entrance slits for the 3 spectrometers were carefully co-aligned during the optical bench assembly to ensure that all 3 spectrometers would share a common bore site. After the instrument vibration test, an optical component in the 1.61 µm CO2 channel shifted, introducing a ~70 arcsec shift in the bore site of that channel. The root cause of the misalignment was traced to a specific step in the optical bench assembly process. While it was not possible to correct this misalignment for the OCO instrument, a second vibration test was performed to ensure that no further movement would occur, and the science algorithms were modified to accommodate the pointing offset.

For the OCO-2 instrument, the optical bench assembly process has been modified to avoid this problem. This modification will be verified by performing a “seating vibration” followed by an alignment test prior to full optical bench integration.

The OCO instrument used Teledyne mercury cadmium telluride (HgCdTe) FPAs in the 1.61 and 2.06 µm CO2 channels and a silicon, HyViSiTM FPA in the O2 A-band channel. All 3 FPAs used Teledyne HAWAII-1RGTM read-out integrated circuits, so that a common design could be used for their control and readout electronics design.

New FPAs are needed for the OCO-2 instrument for two reasons.

- First, there were not enough high quality spare HgCdTe FPAs from the OCO instrument, to provide flight and flight spare FPAs for the CO2 channels in the OCO-2 instrument.

- Second, there was a strong desire to mitigate the residual image artifacts discovered in the HyViSi FPA during the OCO pre-flight instrument testing. The substrate-removed HgCdTe HAWAII-1RG FPA’s from Teledyne, like those recently flight qualified for the Hubble Space Telescope Wide Field Camera-3 (WF3), could address both of these issues. These FPA’s use the same electrical, thermal, and mechanical interfaces as those on the OCO instrument, minimizing the design changes needed for their accommodation. They also have slightly lower read noise than those used for the OCO instrument. In addition, because these FPAs are sensitive to the wavelengths sampled by the A-band as well as those sampled by the CO2 channels, it might be possible to use these FPA’s in all 3 channels. This both reduces risk and provides an approach for mitigating the HyViSi residual image issues, because the HgCdTe FPAs show no evidence of this problem.

Because of their higher dark currents, the HgCdTe FPAs in the two CO2 channels on OCO were maintained below 120 K, while silicon FPA in the O2 A-band channel was cooled to < 180 K. For the OCO-2 instrument, the cryolinks to the FPA’s have been redesigned to maintain all three FPA’s at < 120 K. This change preserves the option of using substrate-removed HgCdTe FPA’s in all three channels. It may also facilitate the use of an existing HyViSi FPA in the A-band channel. Recent laboratory tests show that operating the A-Band HyViSi FPA at 120 K, rather than 180 K reduces the amplitude of the residual image anomaly to almost undetectable levels.

New cryocooler: To cool its FPAs, the OCO instrument, used a pulse tube cryocooler of NGST (Northrup Grumman Space Technology) that was thermally coupled to an external radiator though variable conductance heat pipes. This cryocooler was the flight spare from the EOS Aura TES (Tropospheric Emission Spectrometer) project, and the last of its kind.

A different cryocooler was therefore needed for the OCO-2 instrument. A single-stage version of the NGST pulse tube cryocooler used by the NOAA ABI (Advanced Baseline Imager) projected for GOES-R (Geostationary Operational Environmental Satellite – R) was adopted to minimize the changes to the instrument’s thermal and electrical interfaces. This cryocooler, referred to as CSS (Cryogenic Subsystem), is slightly smaller and more efficient than the one used by the OCO instrument, but did require changes in the cryocooler electronics and the heat pipes.


Figure 8: Photo of the pulse tube cryocooler (image credit: NASA/JPL)


Figure 9: Photo of CSS integrated with OBA (Optical Bench Assembly), image credit: NASA/JPL


3 co-boresighted, high resolution, imaging grating spectrometers

Spectral bands

O2: 0.765 µm A-band
CO2: 1.61 µm band
CO2: 206 µm band

Resolving power

> 20,000

Fast optics

f/1.8 , high SNR

Swath (FOV of 14 mrad in cross-track)

- 10.6 km at nadir (defined by the slit width for a 705 km orbit)
- 8 cross-track footprints @ 3 Hz
- 1.29 km x 2.25 km footprint at nadir

Instrument mass, power

140 kg, ~ 105 W

Changes from OCO instrument

- New FPAs
- New cryocooler

Table 4: Overview of OCO-2 instrument parameters


Observation modes:

For normal science operations, the spacecraft bus orients the instrument to collect science data in Nadir, Glint, and Target modes. The varies modes optimize the sensitivity and accuracy of the observations for specific applications.

1) Nadir mode: the instrument views the ground directly below the spacecraft (local nadir). Observations are made whenever the solar zenith angle is < 85º.

2) Glint mode: the instrument views the location where sunlight is directly reflected on the Earth's surface (the OCO-2 is pointed towards the bright spot where solar radiation is specularly reflected from the surface). The glint mode enhances the instrument's ability to acquire highly accurate measurements, particularly over the ocean. Also, the Glint measurements over the ocean should provide much higher SNR values. Glint soundings are being collected at all latitudes - whenever the local solar zenith angle is < 75º. OCO-2 switches from Nadir to Glint modes on alternate 16-day global ground-track repeat cycles so that the entire Earth is mapped in each mode on roughly monthly time scales.

3) Target mode: the instrument views a specified surface target (calibration site) continuously as the satellite passes overhead (max pass duration up to 9 minutes). The Target mode provides the capability to collect a large number of measurements over sites where alternative ground-based and airborne instruments also measure atmospheric CO2.

- The instrument collects up to 12000 soundings during a single 9-minute overpass at surface observation zenith angles between 0 and 75º

- A small oscillation can be superimposed on pointing to scan the spectrometer bore site across the target as the observatory flies overhead, imaging a 15 x 30 km area: ideal for mapping point sources within cities.


Figure 10: Schematic view of the observation modes (image credit: NASA/JPL, Ref. 17)

Naturally, the size of the footprint increases when observations are made in Glint or in Target modes (that are different from nadir).

For OCO, the nominal plan was to switch from Nadir to Glint observations on alternate 16-day global ground-track repeat cycles so that the entire Earth is mapped in each mode every 32 days. A similar approach has been adopted for OCO-2. Comparisons between Nadir and Glint observations will provide opportunities to identify and correct for biases introduced by the viewing geometry. Target observation will be acquired over an OCO-2 validation site roughly once each day.

The same data sampling rate is used for Nadir, Glint, and Target observations. In each mode, the instrument can collect up to 8 soundings over its 0.8º wide swath every 0.333 s. For nadir observations from a 705 km orbit, traveling at ~7 km/second, the 0.333 s frame rate yields surface footprints with down-track dimensions < 2.25 km. The cross-track dimension of the swath depends on the orientation of the slit with respect to the orbit path, which changes as the spacecraft travels from south to north along its orbit track.

Near the polar terminators, when the spectrometer slit is oriented perpendicular to the orbit track, the cross-track swath for nadir observations is ~10.5 km wide. At the sub-solar latitude, where the spectrometer slit is almost perpendicular to the orbit track, the cross-track dimension of the swath is limited to the projected width of the slit, which is only about 0.1 km wide at nadir.


Figure 11: Variation of OCO-2 footprint size and orientation during an orbit (image credit: NASA/JPL)


OCO-2 calibration:

Pre-flight calibration: The instrument consists of three imaging spectrometers, one for each band (Figure 12) . The radiometric calibration process requires characterizing both the dark current level and gain coefficients of each instrumental channel. Light is directed into the spectrometers through a common telescope and a series of beam splitters and reimagers. Just before the incoming light enters each spectrometer, a linear polarizer selects the polarization vector parallel to the entrance slit. Each spectrometer works in the first order and uses a flat holographic grating. 18)

At each spectrometer’s focus, an area array collects the spectrum. As is typical in imaging spectrometers, one dimension measures the field angles along the slit, and the other dimension measures the different wavelengths.


Figure 12: Schematic view of the spectrometer optics chain as well as the integrating sphere for the radiometric ground test. Lamp D is external to the integrating sphere, and its brightness is controlled by an adjustable slit (image credit: NASA/JPL)


Calibration approach: The OCO-2 instrument focal planes will record the brightness of the incident spectral radiances as raw data numbers (DN). Data numbers are measures without units. For the OCO-2 mission, data numbers that represent spectral radiances range from 0 to 216. The OCO-2 Ground Data System will be responsible for the conversion of these data numbers into wavelength-dependent measurements that are expressed in meaningful physical units. The physical units that the OCO-2 mission will use for spectral radiance are photons per square meter per steradian per second. Radiometric calibration is the process that collects and applies the parameters needed to convert the instrument output into physical measures. The radiance measure generated by the calibration process will be the critical component of the OCO-2 Level 1B Product. 19)

A dedicated team will oversee the radiometric calibration process for the entire OCO-2 mission. This team will maintain the algorithms and update the parameters required to generate accurate radiances. Ongoing calibration exercises during the space operations of OCO-2 will ensure, that the mission obtains bias-free radiance measurements. Accurate radiance measures are crucial to retrieve XCO2 with the precision needed to determine the geographic distribution of CO2 sources and sinks. This aspect of the OCO-2 mission is vital as CO2 sources and sinks must be inferred from small (<2%) spatial variations in XCO2. OCO-2 will have a precision of <0.3% (1 ppm), thus allowing for the quantification of CO2 sources and sinks.

The calibration team will characterize the instrument on the ground before Observatory launch. The characterization exercise will yield the initial set of parameters required to convert instrument data numbers into incident radiances. Once the instrument begins to operate in flight, its behavior will change due exposure to the space environment. For the remainder of the mission, the calibration team will use on-board calibration capabilities to track instrument behavior, and modify the calibration parameters to ensure accurate assessment of instrument measure.

On orbit calibration: The OCO-2 instrument employs an OBC (On Board Calibrator) to detect changes in the instrument gain and wavelength response. While the spacecraft flies over the dark side of the Earth, the instrument will automatically collect calibration data using the OBC. Unlike the OCO-2 science data, clear sky conditions will not be required to acquire this data. The mission will regularly perform four types of calibration using the OBC. Each calibration generates a unique data collection, which include:

Cal_solar data: To collect these data, the instrument will deploy an attenuation screen in front of the telescope. The Observatory will point the instrument telescope at the Sun while the instrument line of sight is above the Earth's atmosphere. The radiances and the wavelengths of the spectral lines in the solar spectrum are well established. Thus, radiances recorded in the Cal_solar data will provide a means to calibrate the absolute instrument response as well as relative instrument response among the three OCO-2 spectrometers. The wavelengths where radiances appear in the Cal_solar data will also provide a means to calibrate the spectral wavelength associated with each spectral sample.

Cal_limb data: These data are an extension of the Cal_solar data. The Observatory will acquire both data sets in sequence. Acquisition of the Cal_solar data will immediately precede the Cal_limb data. The instrument attenuation screen will remain deployed and the instrument telescope will continue to view the Sun. However, as the Observatory orbit progresses, the instrument's line of sight will pass through the Earth's atmosphere. Thus, Cal_limb spectra will contain both Solar absorption lines as well as absorption lines that are characteristic of the atmosphere's chemical content.


Figure 13: Overview of the OBC (highlighted blue circle) that overlays the instrument telescope/collimator assembly. (image credit: NASA/JPL)

Cal_dark data: The mission will employ two means to collect these data. Either the instrument will view the dark ocean at night, or the instrument will apply a cover to the viewing telescope. Cal_dark data will specify the focal plane response for a totally dark scene. Thus, these values will specify the "zero point" on the radiance scale. Once a calibration is applied, measurements that are equivalent to the "zero point" will indicate no incident light.

Cal_dark data will be collected at two points on the night side of the orbit. One set of Cal_dark data will always collected at the same relative location of the Observatory orbit relative to the day-night terminator. These data monitor long-term drift of the zero point. A second set of Cal_dark data will be collected at different locations over the night side of the orbit. These data will monitor shifts in the zero-point offset that are associated with changes in instrument or spacecraft temperature.

• Cal_lamp data. To collect these data, the instrument will turn on one of three small light bulbs. Light from the bulb will illuminate a reflector. The reflector will diffuse the light to produce a uniform field that is directed into the instrument telescope. Since the spatial and spectral distribution from these bulbs is uniform and well known, Cal_lamp data will provide the "flat fields" that are used to define the relative radiometric response for each detector on the focal plane.

• Vicarious Calibration (VC) will employ precise in situ measurements collected at the Earth's surface to estimate the solar radiation field at the top of the Earth's atmosphere. The calibration team will compare these data with measurements acquired by the Observatory. The comparisons will yield a correction table. Application of the correction table will force OCO-2 measurements to conform to the Vicarious Calibration experiment. This adjustment will provide both an absolute and channel-relative calibration for OCO-2 data products.

• "Flat fielding" will employ Earth scene statistics which will be collected from sets of Nadir Mode data. These statistics will provide a means to verify the sample-relative-gain coefficients generated by the On Board Calibrator. A consistent decrease in the radiance for one sample, as compared to a neighbor sample at the same wavelength in the same spectrum, will indicate an error in the calibration process. When systematic effects are detected, the team will apply the Earth scene statistics to update the characterization of the On Board Calibrator.

• Cal_doppler observations will view the Sun over one entire day side of the orbit. These data will provide measurements of the solar spectrum over the full range of Doppler shifts that the Observatory encounters. These data will also provide a means to apply Doppler corrections to instrument wavelength measurements.


Figure 14: Routine and special calibration observations ensure on-orbit performance (image credit: NASA/JPL)


Ground truth:

To ensure accuracy, the spaceborne CO2 estimates are validated through comparisons with near-simultaneous measurements of CO2 acquired by ground-based Fourier Transform Spectrometers in TCCON (Total Carbon Column Observing Network). This network currently includes over a dozen stations, distributed over a range of latitudes spanning Lauder, New Zealand and Ny Alesund, Norway, and is continuing to grow. To relate TCCON measurements to the WMO CO2 standard, aircraft observations have been collected over several stations, using the same in situ CO2 measurement approaches used to define that standard. OCO-2 will target a TCCON site as often as once each day, acquiring thousands of measurements as it flies overhead. These measurements will be analyzed to reduce biases below 0.1% (0.3 ppm) at these sites. The spaceborne CO2 estimates will be further validated through comparisons with CO2 and surface pressure measurements from ground based sites with the aid of data assimilation models to provide a more complete global assessment of measurement accuracy (Ref. 3).


Ground system:

The OCO ground system is designed around communicating with the observatory once per day for downlink passes. The majority of the on orbit operations repeat in 16 day cycles (oscillating between glint and nadir, orbits). The target observations happen daily, and require daily uplinks. These targets are chosen by the validation team the day prior and communicated to the Orbital Mission Operations Team. Following these target selections, the appropriate commands are sent to the spacecraft. The high volume of science data returned requires a highly automated science & data processing system with a high level of coordination between the science team at JPL and the Operations teams at OSC & GSFC. The OCO-2 ground system is designed to facilitate the coordination of multiple teams around a daily data uplink and downlink (Figure 15).


Figure 15: Architecture of the OCO-2 ground segment (image credit: NASA/JPL, Ref. 5)

1) “OCO-2 home page,” NASA/JPL, URL:

2) Debra Werner, “NASA Gearing Up to Replace Orbiting Carbon Observatory,” Space News, Feb. 22, 2010, p. 12

3) David Crisp, “Measuring CO2 from Space: The NASA Orbiting Carbon Observatory-2,” Proceedings of the 61st IAC (International Astronautical Congress), Prague, Czech Republic, Sept. 27-Oct. 1, 2010, IAC-10.B1.6.2

4) David Crisp, “The Orbiting Carbon Observatory-2 (OCO-2 Mission,” AIAA Space 2010 Conference & Exposition: 'Future Earth Science Missions and Enabling Activities,' Aug. 30 to Sept. 2, 2010, Anaheim CA, USA, URL:

5) Annmarie Eldering, Benjamin Solish, Peter Kahn, Stacey Boland, David Crisp, Michael Gunson, “High Precision Atmospheric CO2 Measurements from Space: The Design and Implementation of OCO-2,” Proceedings of the 2012 IEEE Aerospace Conference, Big Sky, Montana, USA, March 3-10, 2012

6) Ralph R. Basilio, Thomas R. Livermore, Y. Janet Shen, H. Randy Pollock, “The quest for an OCO (Orbiting Carbon Observatory) re-flight,” Proceedings of the SPIE Remote Sensing Conference, Toulouse, France, Vol. 7826, Sept. 20-23, 2010, paper: 7827-10, 'Remote Sensing of Clouds and the Atmosphere XV,' edited by Richard H. Picard, Klaus Schäfer, Adolfo Comeron, doi: 10.1117/12.867042

7) “Orbiting Carbon Observatory-2 (OCO-2),” Orbital Fact Sheet, URL:

8) “OCO-2 Spacecraft,” URL:

9) Joshua Buck, George H. Diller, “NASA Selects Launch Services Contract for Three Missions,” NASA, July 16, 2012, URL:

10) Stephen Clark, “Carbon-sniffing satellite faces one-year delay,” Spaceflight Now, Feb. 10, 2012, URL:

11) Michael Curie, George H. Diller, “NASA Awards Launch Services Contract for OCO-2 Mission,” June 22, 2010, URL:

12) Randy Pollock, Robert E. Haring, James R. Holden, Dean L. Johnson, Andrea Kapitanoff, David Mohlman, Charles Phillips, David Randall, David Rechsteiner, Jose Rivera, Jose I. Rodriguez, Mark A. Schwochert, Brian M. Sutin, “The Orbiting Carbon Observatory Instrument: Performance of the OCO Instrument and Plans for the OCO-2 Instrument,” Proceedings of the SPIE Remote Sensing Conference, Toulouse, France, Vol. 7826, Sept. 20-23, 2010, paper: 7826-28, 'Sensors, Systems, and Next-Generation Satellites XIV,' edited by Roland Meynart, Steven P. Neeck, Haruhisa Shimoda, doi: 10.1117/12.865243

13) M. R. Gunson, A. Eldering, D. Crisp, C.E. Miller, and the OCO-2/ACOS Team, “Progress in Remote Sensing of Carbon Dioxide from SpaceThe ACOS Project,” 39th NOAA ESRL Global Monitoring Annual Conference 2011, Boulder, CO, USA, May 17-18, 2011, URL:

14) David Crisp, “The NASA OCO-2 CO2 directed satellite mission,” Satellite Hyperspectral Sensor Workshop, March 30, 2011, URL:

15) David Crisp, “Measuring CO2 from Space: The NASA Orbiting Carbon Observatory¿2 (OCO¿2),” Poster, 2011, URL:,%20David.pdf

16) “OCO-2 Instrument,” NASA/JPL, URL:

17) David Crisp, “Status of the OCO-2 Mission,” The GOSAT Workshop 2012 - Towards GOSAT-2 Mission, Feb. 29 - March 2, 2012, Tokyo, Japan

18) Christopher W. O’Dell, Jason O. Day, Randy Pollock, Carol J. Bruegge, Denis M. O’Brien, Rebecca Castano, Irina Tkatcheva, Charles E. Miller, David Crisp, “Preflight Radiometric Calibration of the Orbiting Carbon Observatory,” IEEE Transaction on Groscience and Remote Sensing, Vol. 49, Issue 7, July 2011, pp. 2793-2801, URL:

19) “Calibration Overview,” NASA/JPL, URL:

The information compiled and edited in this article was provided by Herbert J. Kramer from his documentation of: ”Observation of the Earth and Its Environment: Survey of Missions and Sensors” (Springer Verlag) as well as many other sources after the publication of the 4th edition in 2002. - Comments and corrections to this article are always welcome for further updates.