10 likes | 76 Views
252 nm. 306 nm. Solar. Earth. Direct Lamp View. Lamp Diffuser View. Ratio. Entrance Aperture. Entrance Aperture. SBUV/2 Wavelengths. Earth View. Solar Diffuser View. Entrance Aperture. Entrance Aperture. Sun. Diffuser. Earth. Hg Lamp. In-band error.
E N D
252 nm 306 nm Solar Earth Direct Lamp View Lamp Diffuser View Ratio Entrance Aperture Entrance Aperture SBUV/2 Wavelengths Earth View Solar Diffuser View Entrance Aperture Entrance Aperture Sun Diffuser Earth Hg Lamp In-band error 280 nm measurements OOBR Corrected Techniques for Calibration, Characterization, and Validation of Ozone Instruments Measuring Back-Scattered Radiances L. Flynn1 (Government Principal Investigator), M. DeLand2, L-K Huang2, S. Taylor2, P. Bhartia3, E. Beach1, Y. Li1, S. Kondragunta1, T. Beck1 1NOAA NESDIS, 2SSAI, 3NASA GSFC Requirement: Develop an integrated global observation and data management system for routine delivery of information on the current state of the climate. Science:Can we create sufficiently accurate time series of global ozone profiles to monitor long-term changes in the ozone for climate change & atmospheric composition studies? Benefit: Achieve faster availability and better accuracy for products in both near real time and climate applications and models. Introduction Surface Reflectivity This poster describes a variety of hard and soft calibration techniques as applied for verification and analysis of in-orbit Solar Backscatter Ultraviolet (SBUV/2) instrument performance. These techniques are used to characterize, estimate and track measurement calibration, wavelength scales, bandpasses, and stray light. Internal consistency checks and external comparisons are demonstrated as applied to validate radiance and irradiance measurements, and total and profile ozone retrievals. The goal is to monitor changes in stratospheric ozone (total column and vertical profile) over multi-decade timescales. This requires accurate data from individual instrument to improve the knowledge of absolute calibration and time-dependent changes. The hard calibration system is augmented by vicarious calibration using analysis and monitoring of the Earth albedo for reflectivity channels. Antarctica and Greenland have high surface reflectivity and year-to-year stability. This allows us to evaluate the absolute calibration and time-dependent drift. Solar zenith angle dependence also provides information about linearity. Multiple SBUV and SBUV/2 instruments provide overlapping data sets covering 25 years Solar irradiance measurements are taken over the same wavelength range used for ozone measurements Albedo Calibration Ozone Pair Comparisons The SBUV/2 instruments measure radiance and irradiance with the same optical system. The solar diffuser is the only element not common to both measurements; we must track reflectivity changes to get accurate albedo calibration. An on-board calibration system provides the baseline for long-term instrument characterization. The SBUV/2 retrieval algorithm computes total ozone estimates by using three different wavelength pairs for small path length (low solar zenith angles) and low total column ozone amounts. The “D-pair” (305.8 nm and 312.5 nm) total ozone estimate has good sensitivity to ozone abundance, but low sensitivity to wavelength-dependent calibration errors. To avoid profile shape effects, “D-pair” values are compared to “A-pair”(312.9 nm and 331.1 nm) and “B-pair” (317.5 nm and 331.1 nm) total ozone estimates at equatorial latitudes when the Solar Zenith Angles are less than 60o. To track the B-pair calibration we assume the D-pair is accurate and examine the differences. The results provide a correction for the B-pair calibration (317.5 nm). An on-board calibration system tracks diffuser reflectivity changes Solar irradiance agrees with reference spectrum within calibration uncertainty over the 200 nm to 400 nm range Comparisons between B-pair and D-pair are used to estimate calibration adjustments for total ozone channels Measurements at lamp or solar lines track the wavelength scale and monitor the bandpass shape Ozone absorption cross sections in the UV for total ozone channel Data SZA<60 Zonal mean initial residuals for low latitudes for 2006 show relative variations among the three operational SBUV/2 instruments Reflectivities at Hg-lamp emission lines help to characterize wavelength-dependent diffuser changes In-flight measurements over a range of angles are used to “bootstrap” diffuser goniometry Analysis of Residuals Measurement residuals – defined as the differences between the measured radiances and those calculated by using a forward model and an ozone profile – are used to track changes between instruments or between channels for a single instrument. Both initial residuals (with respect to a priori ozone profiles) or final residuals (with respect to the ozone profile retrieval algorithm estimate) are used. The former are most useful for comparing instruments. The latter are most useful to identify inter-channel changes. The changes can represent calibration drift or instrument performance or geophysical changes. Ascending/Descending Consistency Stray Light Characterization Notice that 252 nm for 30 deg. (top dashed line) is very similar to 273 nm for 70 deg. (second dotted line) The SBUV/2 instruments make measurements of the sun-lit Earth. At certain times of the year, the same latitude is observed twice, on both the ascending and descending portions of an orbit. The solar zenith angles for the two observations differ. One can find cases where the contribution function for an SBUV/2 channel at the larger solar zenith angle is very similar to the contribution function for another, shorter, SBUV/2 channel at the smaller solar zenith angle. This allows an internal consistency check of the two channels’ relative calibration. The Figure at the right shows the penetration of photons into the atmosphere for 252, 273, 283, 288, 292, 298, 302 and 306-nm channels from top to bottom respectively. The SBUV/2 instruments experience both in-band and out-of-band stray light. The out-of-band stray light is believed to come from internal scattering due to dust and optical roughness and imperfections. Some recent SBUV/2 instruments have experienced in-band stray light. This only occurs at high SZAs when the sun in on a specific side of the satellite. The signal contamination begins above 75 SZA, and continues even on the night-side, ending when the spacecraft enters the Earth shadow at 117 SZA. Analysis using Earth radiance measurements across the 280 nm solar feature to construct a Mg II Index, with and without out-of-band corrections, demonstrate that this additional error has the same spectral dependence as the solar signal, that is, it is in-band stray light. Given its presence on the night-side it is probably sunlight scattered by some part of the instrument or spacecraft. Absolute Adjustments A variety of techniques have been used with the SBUV(/2) instruments in retrospective characterization for reprocessing. NOAA-11 SBUV/2: Identify coincident measurements with SSBUV instrument from multiple flights. Nimbus-7 SBUV: Use overlap with adjusted NOAA-11 data to define initial changes. Reflectivity over ice indicated need to adjust calibration at non-absorbing wavelengths. NOAA-16 SBUV/2: Uniform shift to pre-launch calibration (+5.7%) determined from snow/ice radiance comparisons. Wavelength-dependent variations are minimal. NOAA-16 comparisons: Use microwave, LIDAR data for SBUV/2 wavelengths corresponding to useful altitudes of external data. Validation tests for microwave results are not sensitive to derived linear wavelength dependence. NOAA-9 SBUV/2: Normalize to NOAA-11 in 1993, when both instruments observe at similar solar zenith angles. (not shown) CONCLUSIONS Measurements of sunlight reaching the ground, using the atmosphere as cut-off filter for radiance at wavelength below 290 nm, show that stray light is present at significant level in the shortest five SBUV/2 channels (252, 273, 283, 288 and 292 nm). If the out-of-band stray light is uncorrected, radiance errors produce false correlation of ozone changes with reflectivity changes at 340 nm. The solid line shows an extrapolation using longer wavelength values to predict the true signal at shorter wavelengths. The shorter channels have significant deviations. After adjustment, the final residuals from the Version 8 ozone profile retrieval algorithm are monitored. By using in-orbit correlations between shorter channel measurements and the simultaneous cloud cover radiometer measurements at 380 nm, we can estimate the impact of stray light and construct a single term correction for each channel. Position mode data gives a dense set of measurements for this purpose. Earth-view measurements across the Mg II solar features are also used to identify additive errors with linear wavelength dependence. An example for the 280.7 nm data is shown below. 5% with stray light 380 nm Radiance 380.0 nm Radiances 280.7 nm Radiances without stray light Fit of 280.7 variations with CCR Summary The SBUV/2 instruments have on-board calibration systems to track changes in the diffuser reflectivity. This is supplemented by “soft” calibration techniques using both standard measurements and special science measurements to validate on-board system. This combination of multiple techniques provides long-term calibration and has been used when the onboard calibration system became inoperable. Additional data and analysis techniques are used to track wavelength scale changes, and to identify and estimates stray light contamination. Intercomparisons to measurements and ozone products from other SBUV/2 instruments and from ground and other space-based ozone and UV-measuring systems are used to validate the characterizations. Bhartia, P.K., et al., 1995, “Applications of the Langley Plot Method to the Calibration of SBUV Instrument on Nimbus-7 Satellite,” J. Geophys. Res., 100, 2997-3004. Cebula, R.P., & DeLand, M.T., 1998, “Comparisons of the NOAA-11 SBUV/2, UARS SOLSTICE, and UARS SUSIM Mg II solar activity proxy indexes,” Solar Physics, 177, 117–132. Deland, M.T., et al., NOAA-17 SBUV/2 (Flight Model#6) Activation and Evaluation Phase (A&E) Report, 2002, SSAI Document # SSAI-2015-180-MD-2002-02. Herman, J.R., et al., 1991, “A new self-calibration method applied to TOMS and SBUV backscattered ultraviolet data to determine long-term global ozone change,” J. Geophys. Res. , 96 , 7531-7545. Huang, L.K., et al., 2003, “Determination of NOAA-11 SBUV/2 radiance sensitivity drift based on measurements of polar ice cap radiance,” Int. J. Remote Sensing, 24(2),305-314. References Science Challenges:Changes in instrument behavior and performance in-orbit must be tracked and characterized to produce accurate records. Next Steps:SBUV/2 calibration and validation analysis methods are mature but require continual iterative analysis and reprocessing. Transition Path:Reprocessing occurs withinSTAR.Methods and monitoring plots are undergoing automation and inclusion in the ICVS.