1. Introduction
An objective of the Clouds and the Earth’s Radiant Energy System (CERES) project is to measure the reflected solar radiances and earth-emitted radiances so the radiation fluxes from the “top of the atmosphere” (TOA) can be computed, as was done by the Earth Radiation Budget Experiment (ERBE), although with greater accuracy (Wielicki et al. 1996). An additional objective of the CERES project is to use these measurements together with measurements from other instruments to compute the surface radiation fluxes and the radiation fluxes within the atmosphere. The use of measurements smaller than the CERES pixel has led to the need to know the point response function (PRF) of the CERES radiometers (Wielicki et al. 1998). Consequently, the CERES PRF has been analyzed by Smith (1994). Also, measurements were made during calibration testing on the ground from which the PRF was computed for the various channels of each of the instruments (Paden et al. 1997, 1998). Lunar observations provide an opportunity to validate the PRF in orbit.
The CERES scanning radiometers were operated to make measurements of the moon with the objective of utilizing the full moon as a quasi-point source. The primary goal was to validate the point response functions of the channels under near-steady-state conditions during raster scans of the moon. These data are also used to demonstrate the feasibility of validating the pointing accuracy of the CERES channels.
To use the data to make maps of radiation fluxes at TOA on a 1° grid, a pointing knowledge of several kilometers is adequate. However, to use CERES measurements with higher-resolution data from the Moderate Resolution Imaging Spectroradiometer (MODIS), it is necessary to have geolocations of the pixels that approach the size and spacing of the MODIS pixels. Thus, the pointing accuracy must be validated to a few kilometers. This level of accuracy was validated by Currey and Smith (1998) by use of coastline detection for the protoflight model of CERES, aboard the Tropical Rainfall Measuring Mission and used for the CERES FM-1 and FM-2 instruments aboard the Terra spacecraft and the FM-3 and FM-4 instruments aboard the Aqua spacecraft (Smith et al. 2009). An alternative method of validating pointing accuracy is provided by lunar observations.
2. Experiment design
The field of view (FOV) of a CERES radiometer is determined by a field stop in front of the detector. This field stop is a hexagon formed by truncating a square diagonally as shown in Fig. 1. The FOV is 1.3° wide in the scan direction and 2.6° across the corners. The telescope has a blur circle with a radius of 0.16° (Smith et al. 2001). Nominally, the detector has a uniform response over the portion within the FOV. The contours in Fig. 1 show the design static PRF of each channel as computed by a ray trace program (Haeffelin et al. 1997). The static PRF is uniform over the FOV except at the edges, where the blur circle results in a ramp at the edges. Also, there is a small blurring at the edges due to lateral conduction of heat across the thickness of the detector, not shown here. The point spread function (PSF) together with the lateral conduction forms the static PRF. The moon subtends an arc of 0.5° (diameter), so that it may be considered a quasi-point source for the CERES.
The thermistor bolometer detectors have a time response of about 8 ms and the electronic signal conditioning circuit is a 22-Hz four-pole Bessel filter. For earth scanning, the radiometer scans at 63.5° s−1 in elevation angle, thus the time response of the instrument has a strong effect on the PRF (Smith 1994). To validate the alignment of the channels, the instrument is operated during the scan of the moon with the elevation angle fixed and with a slow scan in azimuth of the moon so the time response of the instrument is of minor importance to the PRF in these measurements. This slow scan of the moon permits measurements of the static PRF, which is determined by the optics and the detector, as distinguished from the dynamic PRF, which is the response at a nominal scan rate and includes the time responses of the detector and electronics.
By combining knowledge of the motion of the moon relative to the spacecraft and the programmability of the CERES instruments, a raster scan of the FOV by the moon was planned. Figure 2 shows the movement of the moon across the FOV of a CERES channel as the instrument scans in azimuth. As the spacecraft moves along its orbit, it rotates at a rate of 3.67° min−1 to keep its earth-facing side aligned with nadir, thereby moving the moon’s location in elevation.
Observations of the moon for this purpose are made at the near-full moon, with the lunar phase angle less than 7° to get a large signal but greater than 4° to avoid the surge of lunar radiance. These criteria permit making measurements 3 days before and after the full moon.
The ascending node of Terra is at 2230 LT, so as the Terra spacecraft comes over the Arctic region the moon is in a good position to be viewed when the instrument scans to make a space look behind the spacecraft. Looking in the direction opposite the velocity vector minimizes any contamination due to the ram effect of looking forward. The ascending node of the Aqua spacecraft is at 1330 LT, so that it views the full moon behind the Aqua spacecraft as it passes Antarctica. Good conditions occur once per month for each spacecraft, and these tests are performed quarterly.
If the spacecraft were in true polar orbit, the moon would not move in azimuth as the spacecraft moved along its orbit. However, the orbit inclination is 81° retrograde, so the lunar track has a slow apparent movement in azimuth as seen by the instrument. The instrument is fixed at an elevation angle of 8° as the instrument scans in azimuth at 4° s−1, and the moon moves in elevation due to the rotation of the spacecraft. The time here is denoted as t = 1 min. After this raster scan is complete, the instrument is repositioned in elevation to an angle of 13° and t = 3 min and the raster is repeated. Finally, another raster scan is performed at an elevation angle of 18° at t = 5 min. The time when the instrument is not looking at the moon serves as a space clamp, or zero radiance measurement.
3. Results
Lunar observations are presented for FM-3 aboard the Aqua spacecraft on 16 February 2003. Next, the measurements are assembled to produce a map of the FM-3 PSF. Similar results are shown for all four CERES instruments aboard the Terra and Aqua spacecraft. Anomalies are revealed in the window channel of FM-2. The cause of this departure from nominal response is then discussed.
Figure 3 shows the computed position of the moon in terms of elevation and azimuth from the CERES optical axis of FM-3 during the scans at three elevation positions for one orbit. During the scan in azimuth, there is a small change of elevation due to the motion the spacecraft along the orbit. The center of the moon for the first and last scans is just outside the FOV.
Figure 4 shows the output for the total channel in counts as a function of elevation for the successive scans. Each peak corresponds to a scan in azimuth and is a cross section of the point response function. Figure 5 shows the measurements from the total channel as a function of azimuth angle. The results in Figs. 4 and 5 depend largely on the finite size of the moon and the blur circle of the optics.
a. Effects of the finite size of the moon and optical blur
We assume that the moon is a disc with a radius of 0.25° and uniform radiance of unity. The image of the moon at the aperture has a radiance of m(x, y) = ∫∫DB(x′ − x, y′ − y) dx′ dy′, where B(x, y) describes the blur, and the domain of integration D is the disc of the moon. The blur circle has a radius of 0.16°, so the image has a uniform radiance within a radius r = 0.25 − 0.16 = 0.09 and has 0 radiance beyond r = 0.25 + 0.16 = 0.41. We approximate the radiance of the image by using a cubic expression between r = 0.09 and r = 0.41, so the radiance of the moon’s image on the detector varies with distance from the center of the moon’s image as in Fig. 6. The measurement of the moon as its image moves across the FOV is the integral of the radiance described by Fig. 6 over the part of the image within the FOV and is shown in Fig. 7 as a function of the distance of the center of the moon’s image from the edge of the FOV. The integrated radiance over the 0.25° radius image before blur is π/16 and this is not changed by blur.
The observations of Figs. 4 and 5 have the same shape as the computed response shown by Fig. 6 for the ramp up and ramp down. This curve is repeatable and precise, and furthermore is steep so the location of the edge of the FOV can be located accurately by noting the abscissa, where the response is one-half of the maximum value. The resolution is also affected by the sampling. In the x direction, the sampling interval is the azimuthal scan rate of 4° s−1 divided by the data sampling rate of 100 samples s−1, or 0.04°. In the y direction, the sampling interval is the 1.3° width of the FOV divided by the number of scans. For the case shown in Figs. 3, 4, and 5, there were 11 scans for one orbit. With data from three orbits, there were 33 scans in azimuth, giving a spacing of 0.04° in the y direction.
b. Estimation of location errors
The error in the elevation angle can be computed by the use of Figs. 4 and 7. The scan near elevation −0.77° peaks at 0.45 of the “plateau” value, that is, the average of the scan points away from the edges of the FOV. By Fig. 7, a value of 0.45 corresponds to the moon being 0.1° outside the FOV, so the edge is −0.67° rather than −0.65°, implying an error in elevation angle of 0.02°. Similarly, the scan near the 0.67° elevation angle peaks at 0.27 of the plateau value, which corresponds to the moon being 0.15° outside the FOV. This edge of the FOV is computed to be at 0.52° instead of 0.65°, an error of 0.13°. The center of the FOV is thus at −0.07° in the elevation direction.
The error in azimuth can be estimated by the examination of Fig. 5. The profiles of the output of the total channel during azimuthal scans are not symmetric about azimuth = 0, but they are all shifted to the right by approximately 0.2°, which is taken to be the azimuthal location error. Smith et al. (2009) computed errors of 1–2 km using ground scenes. For the Terra and Aqua altitude of 705 km, this corresponds to 0.08°–0.16°. The errors in elevation and azimuth computed for this case are compatible with the ground validation. However, only one case is shown here and more study is needed.
c. Point response functions
Figure 8 shows contour maps developed from the lunar radiance measurements for the PRFs of each of the three channels of all four CERES instruments. The PRFs of the FM-1, FM-3, and FM-4 instruments and the total and shortwave channels of FM-2 are all flat within the FOV, with the same response curve near the edges of the FOV. The PSFs constructed from lunar measurements are as expected from theory (Smith 1994) and ground measurements (Paden et al. 1998, 1999).
The PRF for the window channel of FM-2 has a region of anomalously high response at the right corner, which was not noted prior to launch but was found in the first lunar observations made by FM-2. This high response indicates that the lunar irradiance in this area causes the detector to heat more than in other areas.
Figure 9 is a drawing of the construction of the thermistor bolometer. It consists of a paint layer, which is bonded to a sealing layer, then to the thermistor. The thermistor is in turn bonded to a layer of kapton, which provides thermal impedance for the thermistor. The kapton is bonded to an aluminum base that acts as a heat sink. The region of high sensitivity indicates that the thermistor may have become delaminated from the kapton or the kapton may have become delaminated from the base. Such delamination can happen when the instrument is placed in the vacuum of orbit, due to air bubbles within the layers.
Figure 10 is a schematic of the effects of delamination on the measurement. The left part of Fig. 10 shows the temperature variation through the normal thermistor bolometer. Radiation absorbed in the paint layer increases its temperature, and this heat is conducted to the thermistor. The increase of temperature of the thermistor changes its resistance, which is the measured quantity. The heat is then conducted through the kapton and epoxy layers to the aluminum heat sink that provides a stable temperature. The right part of the figure shows the effect of delamination. If the kapton layer is not in contact with the aluminum heat sink, the heat does not flow through this layer from the thermistor, so the thermistor heats to a higher temperature for a given radiance, thus causing a greater response.
This anomaly of the window channel of FM-2 is not expected to have any impact on the use of the data. The total and shortwave channel data are used with higher-resolution MODIS data to compute scene identifications of the pixels and to compute the radiation fluxes at the top of atmosphere, the surface, and through the atmosphere. For this purpose the PRF must be well known. Fortunately, the window channel is not used in this manner. The internal blackbody, which is used for onboard calibration of the window channel, fills the FOV of the channel so the window channel measurements will give the correct average of window channel radiance over a number of measurements, such as the average over a 1° latitude and longitude grid box.
4. Conclusions
It has been demonstrated that the CERES instruments can be operated to make observations of the moon to validate the point response functions of all of the channels and to validate the pointing accuracy of the instruments. The pointing accuracy of the total channel was found to be within 0.1° for elevation and 0.2° for azimuth for one case, matching the results from validation studies with ground scenes. Additional studies are needed to give a stronger statistical basis for this result.
The point response functions of all channels were nominal, that is, had uniform response of the FOV, with the response falling off at the edges as expected, except for the FM-2 window channel. A region of very high response was found in the point response function of the window channel of the FM-2 instrument. This anomaly is attributed to delamination of the detector flake from the base. Onboard calibration using the internal blackbody will provide measurements that have the correct average. Data from this channel are not used with higher-resolution data from other sources, so the nonuniformity is not significant.
Acknowledgments
The authors gratefully acknowledge the support of the CERES Program by the Earth Science Office of NASA and the Sciences Directorate of the Langley Research Center for contract support of Science Systems Applications, Inc. and National Institute for Aerospace.
REFERENCES
Currey, C., Smith G. L. , and Neely R. , 1998: Evaluation of the clouds and the earth’s radiant energy system (CERES) scanner pointing accuracy based on a coastline detection system. Earth Observing Systems III, W. L. Barnes, Ed., International Society for Optical Engineering (SPIE Proceedings, Vol. 3439), 367, doi:10.1117/12.325642.
Haeffelin, M. P. A., Mahan J. R. , and Priestley K. J. , 1997: Predicted dynamic electrothermal performance of thermistor bolometer radiometers for Earth radiation budget applications. Appl. Opt., 36 , 7129–7142.
Paden, J., Smith G. L. , Lee R. B. III, Pandey D. K. , and Thomas S. , 1997: Reality check: A point response function (PRF) comparison of theory to measurements for the clouds and the Earth’s radiant energy system (CERES) tropical rainfall measuring mission (TRMM) instrument. Visual Information Processing VI, S. K. Park and R. D. Juday, Eds., International Society for Optical Engineering (SPIE Proceedings, Vol. 3074), 109, doi:10.1117/12.280612.
Paden, J., Smith G. L. , Lee R. B. III, Pandey D. K. , Priestley K. J. , Biting H. , Thomas S. , and Wilson R. S. , 1998: Comparisons between point response function measurements and theory for the clouds and the Earth’s radiant energy system (CERES) TRMM and the EOS-AM spacecraft thermistor bolometer sensors. Earth Observing Systems III, W. L. Barnes, Ed., International Society for Optical Engineering (SPIE Proceedings, Vol. 3439), 344, doi:10.1117/12.325640.
Paden, J., Smith G. L. , Lee R. B. III, Pandey D. K. , Priestley K. J. , Thomas S. , and Wilson R. S. , 1999: Point response characteristics for the CERES/EOS-PM, FM3, and FM4 instruments. Earth Observing Systems IV, W. L. Barnes, Ed., International Society for Optical Engineering (SPIE Proceedings, Vol. 3750), 395, doi:10.1117/12.363536.
Smith, G. L., 1994: Effects of time response on point spread function of a scanning radiometer. Appl. Opt., 33 , 7031–7037.
Smith, G. L., Peterson G. L. , Lee R. B. III, and Barkstrom B. R. , 2001: Optical design of the CERES telescope. Earth Observing Systems VI, W. L. Barnes, Ed., International Society for Optical Engineering (SPIE Proceedings, Vol. 4482), 269, doi:10.1117/12.453463.
Smith, G. L., Priestley K. J. , Hess P. C. , and Currey J. C. , 2009: Validation of geolocation of measurements of Clouds and the Earth’s Radiant Energy System (CERES) scanning radiometers aboard three spacecraft. J. Atmos. Oceanic Technol., 26 , 2379–2391.
Wielicki, B. A., Barkstrom B. R. , Harrison E. F. , Lee R. B. III, Smith G. L. , and Cooper J. E. , 1996: Clouds and the Earth’s Radiant Energy System (CERES): An earth observing system experiment. Bull. Amer. Meteor. Soc., 77 , 853–868.
Wielicki, B. A., and Coauthors, 1998: Clouds and the Earth’s Radiant Energy System (CERES): Algorithm overview. IEEE Trans. Geosci. Remote Sens., 36 , 1127–1141.
FOV and nominal PRF of CERES radiometer channels.
Citation: Journal of Atmospheric and Oceanic Technology 27, 6; 10.1175/2010JTECHA1322.1
Experiment concept of movement of the moon across the FOV during raster scan of CERES radiometer.
Citation: Journal of Atmospheric and Oceanic Technology 27, 6; 10.1175/2010JTECHA1322.1
Pattern of scans in terms of elevation and azimuth directions across the moon for the FM-3 total channel on 16 Feb 2003.
Citation: Journal of Atmospheric and Oceanic Technology 27, 6; 10.1175/2010JTECHA1322.1
Output of total channel of FM-3 as a function of elevation angle for successive azimuth scans for lunar observations on 16 Feb 2003.
Citation: Journal of Atmospheric and Oceanic Technology 27, 6; 10.1175/2010JTECHA1322.1
Output of total channel of FM-3 as a function of azimuth scan angle for differing elevation angles.
Citation: Journal of Atmospheric and Oceanic Technology 27, 6; 10.1175/2010JTECHA1322.1
Normalized radiance of image of the moon for uniform moon due to “blur circle” at secondary aperture as function of distance from center of image of moon in degrees.
Citation: Journal of Atmospheric and Oceanic Technology 27, 6; 10.1175/2010JTECHA1322.1
Response of detector as the image of the moon moves into FOV, as a function of distance of the center of the image from the edge of FOV in azimuthal direction in degrees.
Citation: Journal of Atmospheric and Oceanic Technology 27, 6; 10.1175/2010JTECHA1322.1
Contour maps of PRFs of (left) total, (middle) shortwave, and (right) longwave window channel for (top to bottom) FM-1, -2, -3 and -4 from lunar scan measurements.
Citation: Journal of Atmospheric and Oceanic Technology 27, 6; 10.1175/2010JTECHA1322.1
Schematic drawing of the construction of CERES thermistor bolometer detector.
Citation: Journal of Atmospheric and Oceanic Technology 27, 6; 10.1175/2010JTECHA1322.1
Effect of delamination on temperature distribution across detector and detector response.
Citation: Journal of Atmospheric and Oceanic Technology 27, 6; 10.1175/2010JTECHA1322.1
