Jeffrey S. Czapla-Myers

Landsat-8 Operational Land Imager Radiometric Calibration and Stability

The Landsat-8 Operational Land Imager (OLI) was radiometrically calibrated prior to launch in terms of spectral radiance, using an integrating sphere source traceable to National Institute of Standards and Technology (NIST) standards of spectral irradiance. It was calibrated on-orbit in terms of reflectance using diffusers characterized prior to launch using NIST traceable standards. The radiance calibration was performed with an uncertainty of ~3%; the reflectance calibration to an uncertainty of ~2%. On-orbit, multiple calibration techniques indicate that the sensor has been stable to better than 0.3% to date, with the exception of the shortest wavelength band, which has degraded about 1.0%. A transfer to orbit experiment conducted using the OLI’s heliostat-illuminated diffuser suggests that some bands increased in sensitivity on transition to orbit by as much as 5%, with an uncertainty of ~2.5%. On-orbit comparisons to other instruments and vicarious calibration techniques show the radiance (without a transfer to orbit adjustment), and reflectance calibrations generally agree with other instruments and ground measurements to within the uncertainties. Calibration coefficients are provided with the data products to convert to either radiance or reflectance units.

The Ground-Based Absolute Radiometric Calibration of Landsat 8 OLI

This paper presents the vicarious calibration results of Landsat 8 OLI that were obtained using the reflectance-based approach at test sites in Nevada, California, Arizona, and South Dakota, USA. Additional data were obtained using the Radiometric Calibration Test Site, which is a suite of instruments located at Railroad Valley, Nevada, USA. The results for the top-of-atmosphere spectral radiance show an average difference of −2.7, −0.8, 1.5, 2.0, 0.0, 3.6, 5.8, and 0.7% in OLI bands 1–8 as compared to an average of all of the ground-based measurements. The top-of-atmosphere spectral reflectance shows an average difference of 1.6, 1.3, 2.0, 1.9, 0.9, 2.1, 3.1, and 2.1% from the ground-based measurements. Except for OLI band 7, the spectral radiance results are generally within ±5% of the design specifications, and the reflectance results are generally within ±3% of the design specifications. The results from the data collected during the tandem Landsat 7 and 8 flight in March 2013 indicate that ETM+ and OLI agree to each other to within ±2% in similar bands in top-of-atmosphere spectral radiance, and to within ±4% in top-of-atmosphere spectral reflectance.

Absolute radiometric calibration of the RapidEye multispectral imager using the reflectance-based vicarious calibration method

RapidEye AG is a commercial provider of geospatial information products and customized solutions derived from Earth observation image data. The source of the data is the RapidEye constellation consisting of five low-earth-orbit imaging satellites. We describe the rationale, methods, and results of a reflectance-based vicarious calibration campaign that was conducted between April 2009 and May 2010 at Railroad Valley Playa and Ivanpah Playa to determine the on-orbit radiometric accuracy of the RapidEye sensor. In situ surface spectral reflectance measurements of known ground targets and an assessment of the atmospheric conditions above the sites were taken during spacecraft overpasses. The ground data are used as input to a radiative transfer code to compute a band-specific top-of-atmosphere spectral radiance. A comparison of these predicted values based on absolute physical data to the measured at-sensor spectral radiance provide the absolute calibration of the sensor. Initial assessments show that the RapidEye sensor response is within 8% of the predicted values. Outcomes from this campaign are then used to update the calibration parameters in the ground segment processing system. Subsequent verification events confirmed that the measured RapidEye response improved to within 4% of the predictions based on the vicarious calibration method.

Vicarious calibration of Aqua and Terra MODIS

The Moderate Resolution Imaging Spectroradiometer (MODIS) is onboard both the Terra and Aqua platforms. An important aspect of the use of MODIS, and other Earth Science Enterprise sensors, has been the characterization and calibration of the sensors and validation of their data products. The Remote Sensing Group at the University of Arizona has been active in this area through the use of ground- based test sites. This paper presents the results from the reflectance-base approach using the Railroad Valley Playa test site in Nevada for both Aqua and Terra MODIS. The key to the approach is the measurement of surface reflectance over a 1-km2 area of the playa and results from this method shows agreement with both MODIS sensors to better than 5%. Early results indicate that while the two sensors both agree with the ground-based measurements to within the uncertainties of the reflectance-based approach, there were significant differences between the Aqua and Terra MODIS for data prior to September 2002. Recent results indicate that this bias, if any, is now within the uncertainties of the reflectance-based method of calibration.

Landsat-7 ETM+: 12 Years On-Orbit Reflective-Band Radiometric Performance

The Landsat-7 ETM+ sensor has been operating on orbit for more than 12 years, and characterizations of its performance have been ongoing over this period. In general, the radiometric performance of the instrument has been remarkably stable: 1) noise performance has degraded by 2% or less overall, with a few detectors displaying step changes in noise of 2% or less; 2) coherent noise frequencies and magnitudes have generally been stable, though the within-scan amplitude variation of the 20 kHz noise in bands 1 and 8 disappeared with the failure of the scan line corrector and a new similar frequency noise (now about 18 kHz) has appeared in two detectors in band 5 and increased in magnitude with time; 3) bias stability has been better than 0.25 DN out of a normal value of 15 DN in high gain; 4) relative gains, the differences in response between the detectors in the band, have generally changed by 0.1% or less over the mission, with the exception of a few detectors with a step response change of 1% or less; and 5) gain stability averaged across all detectors in a band, which is related to the stability of the absolute calibration, has been more stable than the techniques used to measure it. Due to the inability to confirm changes in the gain (beyond a few detectors that have been corrected back to the band average), ETM+ reflective band data continues to be calibrated with the prelaunch measured gains. In the worst case, some bands may have changed as much as 2% in uncompensated absolute calibration over the 12 years.

Design and calibration of field deployable ground-viewing radiometers

Three improved ground-viewing radiometers were built to support the Radiometric Calibration Test Site (RadCaTS) developed by the Remote Sensing Group (RSG) at the University of Arizona. Improved over previous light-emitting diode based versions, these filter-based radiometers employ seven silicon detectors and one InGaAs detector covering a wavelength range of 400-1550 nm. They are temperature controlled and designed for greater stability and lower noise. The radiometer systems show signal-to-noise ratios of greater than 1000 for all eight channels at typical field calibration signal levels. Predeployment laboratory radiance calibrations using a 1 m spherical integrating source compare well with in situ field calibrations using the solar radiation based calibration method; all bands are within ±2.7% for the case tested.

Radiometric calibration of earth-observing sensors using an automated test site at Railroad Valley, Nevada

The Remote Sensing Group (RSG) at the University of Arizona uses the reflectance-based approach to radiometrically calibrate airborne and spaceborne sensors in the solar-reflective regime. The Radiometric Calibration Test Site (RadCaTS) concept was developed in 2004 to increase the amount of ground-based data collected. RadCaTS provides a methodology to determine the surface reflectance for any arbitrary test site in the absence of ground personnel. It is founded on the reflectance-based approach and has successfully operated at Railroad Valley, Nevada, with a suite of instruments including nadir-viewing multispectral radiometers, a Cimel sun photometer, and a meteorological station. RadCaTS data are currently used by RSG to supplement those collected by on-site personnel. This work presents a description of the RadCaTS automated concept, including the process used to determine surface reflectance and top-of-atmosphere (TOA) spectral radiance. The instrumentation required to measure the surface and atmosphere is introduced, followed by discussions regarding their placement on the 1 km2 site at Railroad Valley and their calibration. Lastly, the RadCaTS results are compared with those obtained from the Landsat 7 Enhanced Thematic Mapper Plus (ETM+) and Terra Moderate Resolution Imaging Spectrometer (MODIS). The average percent difference in TOA spectral radiance is 4.1% between the six bands of ETM+ and RadCaTS and 3.6% between the seven land bands of Terra MODIS and RadCaTS.

Design, calibration, and characterization of a field radiometer using light-emitting diodes as detectors

The Remote Sensing Group at the University of Arizona has developed multispectral ground-viewing radiometers that use light-emitting diodes as detectors. This work describes the optical design, electrical design, and laboratory calibration of a three-channel radiometer that operates in the visible and near-infrared region of the spectrum. The optical and electrical design of the radiometer is introduced, and then the calibration and characterization of the radiometer are described. Laboratory measurements include the spectral responsivity for each channel of the radiometer, the temperature dependence of the total responsivity for each channel, system linearity, field of view, and finally, the absolute radiometric calibration. A solar-radiation-based calibration is used to determine the absolute responsivity.

Early ground-based vicarious calibration results for Landsat 8 OLI

The Operational Land Imager (OLI) is one of two instruments onboard the Landsat 8 platform, which was launched on 11 February 2013 from Vandenberg Air Force Base in California. The multispectral bands of OLI retain the 30-m spatial resolution of Landsat 5 TM and Landsat 7 ETM+, but improvements to the system include 12-bit radiometric resolution, eight multispectral bands in the VNIR and SWIR spectral regions, and one panchromatic band. The earlier TM and ETM+ sensors use a whiskbroom configuration, while OLI uses a pushbroom configuration, which allows it to have a higher signal-to-noise ratio than previous Landsat instruments. This also creates challenges in radiometric calibration due to the large number of detectors on the 14 focal plane modules. Long-term data continuity is a crucial component of the 40-year Landsat series of satellites, and ground-based vicarious calibration has played an important role in ensuring that these sensors remain on the same radiometric scale. This work presents the early ground-based in-flight radiometric calibration of OLI, which was determined using the traditional and well-understood reflectance-based approach, as well as the Radiometric Calibration Test Site (RadCaTS), which is an automated suite of instruments located at Railroad Valley, Nevada.

Temporal, spectral, and spatial study of the automated vicarious calibration test site at Railroad Valley, Nevada

The Remote Sensing Group at the University of Arizona has developed an automated methodology and instrument suite to measure the surface reflectance of the vicarious calibration test site at Railroad Valley, Nevada. Surface reflectance is a critical variable used as one of the inputs into a radiative transfer code to predict the top-of-atmosphere radiance, and inexpensive and robust ground-viewing radiometers have been present at the site since 2004. The goal of the automated approach is to retain RSGs current 2-3% level of uncertainty while increasing the number of data sets collected throughout the year without the need for on-site personnel. A previous study was completed to determine if the number and positions of the four radiometers were adequate to spatially sample the 1-km2 large-footprint site at Railroad Valley. The preliminary study utilized one set of panchromatic data from Digital Globes QuickBird satellite. Results from this one day showed that the positions of the four ground-viewing radiometers adequately sample the site. The work presented here expands in a spectral and temporal sense by using high-spatial-resolution data from Ikonos, QuickBird, and Landsat-7 ETM+ to determine if the locations of the ground-viewing radiometers correctly sample the site. The multispectral capability of these sensors is used to establish if there are any spectral effects, which will also help RSG to determine what spectral bands should be chosen for the new ground-viewing radiometers that are currently in development for the automated test site at Railroad Valley.