 |
 Technical Sessions
Pre–launch NPP Sensor Characterizations
Wednesday, August 27, 2008
- Eugene Kratz – Raytheon Information Systems
ABSTRACT: An automated software module has been developed to support the post launch calibration and validation of the CrIS temperature and water vapor profile science data products from the NASA NPP and NOAA/AF NPOESS. This software module runs daily on the NPOESS Science Investigator-led Processing System (NSIPS) infrastructure and utilizes correlative data such as global operational radiosonde measurements (RAOBS) and forecast data from the National Center for Environmental Prediction Global Forecasting System (NCEP GFS), as well as operational profile retrievals generated by other sensors. The software is currently being tested on the operational profile products from both the Atmospheric Infrared Sounder (AIRS) aboard the NASA AQUA spacecraft and the Infrared Atmospheric Sounding Interferometer (IASI) aboard the Eumetsat METOP-A spacecraft. Bias and RMS differences are computed between the collocated RAOB and forecast data and the AIRS and IASI profile products for a variety of globally stratified categories such as sea, land, clear sky, and partial cloudiness. Results of the inter-comparison reveal that the RMS differences for most of the categories are close to the expected goal accuracies (1K in 1 km layers for the temperature and better than 10%-15% in 2-km layers for the water vapor in the troposphere) of the respective sensors. The software also generates calibration and validation data products such as matchup records which contain all the matched correlative data, the product data, and all the necessary additional data to allow the user to re-generate the profile products with different data assumptions. Quality Control, graphical, and analysis capabilities of the software module will also be demonstrated. |
- Mark Esplin, Gail Bingham, Vladimir Zavyalov, Brandon Campbell, Marc Struthers – USU/Space Dynamics Laboratory
ABSTRACT: The Cross-track Infrared Sounder(CrIS) will provide improved atmospheric moisture and temperature profile measurements that can be used to improve the accuracy of future weather forecasts. The Flight Model One (FM1) sensor will be flown on the NPP spacecraft. Obtaining accurate temperature and water vapor profiles will critically depend on the radiometric calibration of the sensor. An extensive series of ground tests are being performed to ensure that the CrIS FM1 sensor is performing correctly and that radiometric and spectral performance of the sensor will be adequate for its planned mission. SDL along with other NPOESS team members are analyzing the test data paying particular attention to the absolute radiometric calibration of the sensor.
Absolute radiometric calibration of a sensor depends on factors such as detector and electronics linearity as well as stability of the optical and electronic systems even if the temperature of the sensor changes. A challenge with the CrIS sensor is that it is a warm instrument with only the detectors being cooled. If not properly accounted for, fluctuation in the instrument background thermal emissions alone would result in fluctuations in the measured earth scene radiances. To minimize this problem, each 8 second cross-scan sweep includes a view of the internal calibration target and a view of cold space.
During thermal vacuum ground testing, sensor linearity is verified by stepping the temperature of a test blackbody through 6 temperatures that roughly cover the range of temperatures seen in the earth’s atmosphere. These measurements are performed at each of three different temperature plateaus that span the expected sensor on orbit temperature. Also during thermal vacuum testing the CrIS sensor will make three simulated orbits where the temperature and the bus voltage will be varied in a way representative of what will happen on orbit. |
- Larrabee Strow, Howard Motteler, Scott Hannon – University of Maryland Baltimore County; David Tobin, Joe Taylor, Lori Borg, Graeme Martin, Hank Revercomb – University of Wisconsin, SSEC
ABSTRACT: The spectral performance of the Cross-track Infrared Sensor (CrIS) during thermal vacuum pre-flight testing will be presented. The CrIS long-wave band instrument lineshape (ILS) was measured using a CO2 laser source for all 9 detectors in the 3x3 focal plane. The spectral performance of all three CrIS bands were tested by recording gas cell spectra at three nominal instrument operating temperatures. These tests validate both the ILS widths and frequency calibration using an internal Ne lamp. |
- David Tobin, Hank Revercomb, Fred Best, Lori Borg, Robert Knuteson, Joe Taylor – University of Wisconsin, SSEC
ABSTRACT: The CrIS Flight Model 1 has recently completed thermal vacuum testing. Here we present analyses of various test data to assess the radiometric linearity and radiometric noise of the sensor. |
- ZiPing (Frank) Sun, Bruce Hauss, Paul Lee – Northrop Grumman Space Technology
ABSTRACT: Based on current orbit assumptions, the NPOESS Preparatory Program (NPP) satellite will be placed in nearly the same orbit plane as that of the NASA “A-Train” satellites, but at a higher orbit altitude (824 km versus 705 km). While the NPP satellite is not an A-Train component, it will form a flying pattern that provides a time window of several hours every 64 hours for cross comparing Sensor Data Record (SDR) and Environmental Data Record (EDR) products from the NPP sensors with those from sensors onboard the A-train satellites. The NPP sensors that can be compared by this process include the Visible/Infrared Imager/Radiometer Suite (VIIRS) with MODIS, the Crosstrack Infrared Sounder (CrIS) with AIRS, the Advanced Technology Microwave Sounder (ATMS) with AMSU, and the Ozone Mapping and Profiler Suite (OMPS) with OMI. Looking ahead to the launch of the National Polar-Orbiting Operational Environmental Satellite System (NPOESS) C1 satellite, the time window overlap with NPP will allow up to several days every 90 days for cross-comparing the VIIRS, CrIS, ATMS, and OMPS sensors on each platform.
As part of the preparation for NPP/NPOESS on-orbit Cal/Val, we are developing the methodology and tools necessary for comparing the measurements from sensors onboard one satellite with those from another. This presentation will illustrate the use of these tools in comparing brightness temperatures from AQUA MODIS with those from NOAA-18 AVHRR/3. We will show that by carefully selecting the time, location and content of the scenes, the brightness temperatures from the two sensors can be compared to an accuracy of better than 0.5K. With this level of accuracy, one can take advantage of satellite cross-comparison as a means for reducing the cost and effort needed in validating new satellite sensors. |
- Glen Jaross, Matthew Kowalewski – SSAI / NASA; Scott Janz – NASA Goddard Space Flight Center
ABSTRACT: Solar diffusers have long been used by satellite sensors measuring in the solar reflective bands in order to maintain calibration and to provide a reference for trace gas absorption measurements. Early ozone sensors such as BUV and TOMS pioneered the technique of normalizing measured Earth radiances by periodic solar irradiance measurements in order to cancel out instrument changes. Apart from bumps along the road related to diffuser degradation, this approach has proved successful and continues to be designed into future sensors.
Following the 1995 launch of GOME on ERS-2, however, it became clear that a fundamental characteristic of some diffusers introduces measurement errors in newer, hyperspectral sensors. Surface-reflecting diffusers, such as the ground aluminum ones used on many UV sensors, cause interference in the reflected radiation that becomes somewhat coherent for highly pixelated detectors. The result is signal patterns referred to as diffuser features that vary with the solar angle of incidence.
In this talk we will describe what is known about diffuser features and show examples from several instruments including pre-launch test results from OMPS/NPP. We will also describe how the features cause problems for atmospheric retrievals and the various approaches to mitigating the problem. One approach involving a fundamental re-design of solar diffusers appears to have worked quite well, and we encourage its incorporation into new sensors. |
- Carl Fischer – MIT Lincoln Laboratory
ABSTRACT: At the National Polar-orbiting Operational Environmental Satellite System (NPOESS) spacecraft level thermal vacuum configuration, contamination concerns on the Visible/Infrared Imager Radiometer Suite (VIIRS) prevent us from opening the nadir aperture doors.As such, the VIIRS sensor has no available cold target for thermal band calibration. However, spacecraft-level testing is the last test phase before launch, so an aliveness test is highly desirable to detect infant mortality or large unexpected gain drifts.
To assess the feasibility of such a test, we have reviewed sensor level thermal vacuum test data from the VIIRS Engineering Design Unit (EDU) and determined that a two-point aliveness test can be performed using the apparent temperature difference between the onboard blackbody calibrator (OBC) and the solar diffuser (SD) reflective band calibration target. We find that signal levels are sufficient to permit two-point vicarious calibration between these two targets. Radiance of the solar diffuser is calculated from thermistors in the sensor scan cavity, based on the assumption that the Spectralon target reflects light from the cavity and is at approximately the same temperature. However, the SD target has large solid angle exposure to unknown temperatures outside the sensor, a likely contributor to observed systematic errors.
Systematic errors in our calculation of the SD radiance approach 25%, but are very repeatable both within a given band and between bands of similar spectral response. Therefore, we expect that this approach will be adequate for the purposes of detecting large changes in emissive band detector response during space-craft level thermal-vacuum testing. However, even with a better model of the SD radiance, significant uncertainties will remain, preventing any accurate emissive calibration with the target. |
- Mau-Song Chou, James McCarthy, John Donovan – Northrop Grumman Space Technology
ABSTRACT: Northrop Grumman Space Technology (NGST) is developing the National Polar-Orbiting Operational Environmental Satellite System (NPOESS) for the tri-agency Integrated Project Office for NASA, NOAA and the DoD.A key sensor, the Visible Infrared Imaging Radiometer Suite (VIIRS) is in development by Raytheon to be deployed on NPOESS. Significant crosstalk and ghosting were observed during the ground tests of VIIRS Engineering Development Unit (EDU) and Flight unit 1 (F1) sensors. In order to model sensor behavior to make sensor data records (SDR) and environmental data records (EDR) performance impact assessments, we have constructed maps of sensor performance with respect to: static electrical crosstalk, optical, dynamic crosstalk, ghosting, and total scattering. The methodologies for construction of these maps will be discussed.
Point-to-point static electrical crosstalk map for F1 is constructed based on sensor-level crosstalk test (FP-13) static collects. Crosstalk influence coefficients are expressed as functions of the number of detectors the receiver is away from the sender detector (i.e. delta detector number). Linearity for static electrical crosstalk is assessed for the dual-gain band with sender and receiver at low gain.
Optical crosstalk map for F1 integrated focal assembly (IFA) was constructed based on filter + detector assembly component-level bench-top test data, corrected for the reflectivity of the dichroic mirror and silver mirrors in the upstream optical train. Near-filed scattering artifacts by test instrument have been identified and discarded. Filter spread function as derived from related bench-top tests shows that the optical crosstalk is limited to only a few detectors adjacent to the sender detector. Uncertainty for optical crosstalk is estimated by RSS of the standard deviation of the mean of the sample numbers with the uncertainty from several errors sources, including the estimate of dark background, instability of light source, spillover, and contamination of instrument near-filed scattering.
Dynamic crosstalk map for F1 is derived based on sensor-level crosstalk test (FP-13) scanning collects. Dynamic crosstalk is found to be linear with the incident radiance, even beyond the saturation of the sender detectors. The observation led to the identification of a root cause in the detector portion of the detector-readout hybrid.
Ghosting map is constructed by combining test data from the scanning slit of the near-field response test (FP-14) and of a special-purpose scanning point-illumination test (STR-547). Absolute ghosting point spread function is derived by normalizing the STR-547 measurements to the ghosting spread function in scan direction, which is the total scattering subtracted by the near field scattering from the FP-14 data. In addition, the NFR in-track point spread function is derived based on the point-to-point crosstalk (FP-13) static collects. A NFR map is the constructed using the NFR point spread function. The total scattering map is the generated by adding the NFR scattering map to the ghosting map. |
- Eugene Waluschka, Peter Fuqua, Sam Pellicori, Tom Mooney, Vijay Murgai, Brian Neff – NASA Goddard Space Flight Center
ABSTRACT: Angle Resolved Scatter (ARS) has recently come to the attention of the earth observing (from space) community as a potential to create significant optical cross talk in the imaging spectro-radiometric instrument. The actual optical cross-talk characteristics may be determined and inferred during ground spectro-radiometric characterization tests. Because of the urgent need to reduce optical cross talk to negligible levels a considerable amount of work, in the past year, has taken place to identify the causes, measure, model and successfully reduce this scatter. We will describe this work here. |
- Patty Pratt, Lushalan Liao, Paul Lee – Northrop Grumman Space Technology; Eugene Waluschka, Hassan Oudrari, Gerhard Meister – NASA/Goddard Space Flight Center; Bruce Guenther – NPOESS Data Products Division, Integrated Program Office
ABSTRACT: To measure the color of the oceans (water-leaving radiance), the polarization of the input light field must properly be characterized. This is done by first computing and correcting for the light path due in part to the solar geometry of the sensor relative to the earth from Rayleigh scattering and then by removing the polarization that was contributed by the sensor. Small errors produced in these steps can attribute false absorption properties to the water resulting in large errors in chlorophyll measurements. The maximum polarization that can be tolerated for any band is 2.5% – 3% as the measured signal for the blue bands is typically only 10% of all contributions to the path radiance. This requires a detailed pre-launch accurate measurement of the response of the sensitivity of the instrument to input polarized light. A characterization uncertainty of 0.5% (one sigma) for both the Degree of Polarization (DoP) and the phase angle specification must also be achieved. Extensive polarization measurements have been recently made and this presentation discusses the current status of the VIIRS polarization characterization. |
- Lushalan Liao – Northrop Grumman Space Technology
ABSTRACT: Using a model-based approach developed by Northrop Grumman Space Technology (NGST), most spatial performance characteristics of the Visible Infrared Imaging Radiometer Suite (VIIRS) to be deployed on the NASA NPP and NOAA/AF NPOESS can be derived. The results of the model-based approach are compared to those from a simpler, less sophisticated, analysis technique. The analysis technique used incorrect assumptions or approximations and resulted in degraded performance. When analyzed correctly, good line spread functions for VIIRS can be derived. Furthermore, the NGST approach revealed extremely near field signals contamination of the order of 3%. This is large in certain S/MWIR and LWIR bands. Most importantly, this approach is directly adaptable for on-orbit use to evaluate spatial performance. It’s has already been demonstrated with MODIS data. |
- John Aldridge, Carl Fischer – MIT Lincoln Laboratory
ABSTRACT: Spectral calibration procedures frequently rely on a tungsten-halogen lamp to serve as a stable irradiance source. For the Visible/Infrared Imager/Radiometer Suite (VIIRS), the original calibration plan called for the use of an integrated tungsten-halogen lamp and linear polarizer to test the instrument for polarization sensitivity. In order to prevent any inherent source polarization from corrupting the measurement, the test lamp and polarizer were fused as a single part and rotated together in an apparatus called the Polarization Source Assembly (PSA).
Previously published studies have shown that the spectral irradiance output of a standard lamp can change if its physical orientation deviates from a normal upright position. This fact was well known, and the VIIRS test plan called for characterization of lamp radiance variations and mitigation through Fourier analysis of the test data. In the end, the PSA characterization method was abandoned in favor of a more traditional test using a stationary lamp, integrating sphere, and a rotating sheet polarizer. We have constructed a testbed capable of measuring lamp radiance variation with position and the variability of this effect from day to day.
In this presentation, we will show the results of several laboratory tests performed on a representative 30 Watt tungsten-halogen lamp, with particular emphasis on spectral irradiance in the 400 nm to 1100 nm region. Tests performed include complete 0-360º studies in lamp orientation under both constant-current and constant-voltage operating modes. Observed effects in these studies include both decreased total irradiance output and decreased relative blue spectral composition when the lamp is utilized outside of the conventional upright position. We also found that measured lamp integrated spectral irradiance is repeatable within a single power-on cycle, but not necessarily between these cycles when tested on a day-to-day basis. |
- Christopher Moeller – University of Wisconsin-Madison
ABSTRACT: The VIIRS FU-1 hardware testing is underway. Ambient environment test data collections have been largely completed with analysis on many of those data sets reaching maturity.The Response vs. Scan (RVS) characterization data was one of those collected during Ambient Phase II in fall 2007 and has been analyzed by Raytheon.
Joint Govt. Team and industry review of the resulting RVS characterization showed that the methodology to capture the SIS100 source output drift and VIIRS gain drift during the test proved to be “not useful” for VIIRS M9 band, a band that is sensitive to water vapor. Using temperature and relative humidity (RH) data collected by a sensor in the ambient laboratory during test data collection, SBRS applied an atmospheric correction to the M9 RVS characterization data. While this improved the overall quadratic fit statistics of the test data set, the M9 drift correction remained not useful. The resulting M9 quadratic fit statistics for RVS were about a factor of eight worse than those of neighboring spectral bands (M8 and M10).
In response to this result, the Govt. Team undertook a task to examine and improve the M9 atmospheric correction using the temperature/RH sensor data as input to the Line-By-Line-Radiative-Transfer-Model (LBLRTM). The RH data in particular showed much short duration structure implying short (1 minute) term fluctuations in RH during test data collection. In the Govt. Team analysis, a bias in the reported time between the test data set and the temperature/RH monitor was adjusted so that the resulting atmospheric correction brought the M9 drift correction in-family with that of bands M8 and M10. The result of this was that the drift correction became “useful” as indicated by an improvement in the quadratic fit statistics of M9 RVS after applying it. The M9 quadratic fit statistics also became closer to those of M8 and M10, though still about a factor of four higher. This effort suggests that applying a careful atmospheric correction to all VIIRS M9 test data sets that include a path through the ambient environment (this is the case for both ambient phase and TVAC phase test data set collection) will result in an improved characterization and reduced uncertainties. |
|
