Absolute Accuracy of Total Solar Irradiance Measurements from

Absolute Accuracy of Total Solar Irradiance Measurements from Space Claus Fröhlich Physikalisch-Meteorlogisches Observatorium Davos, World Radiation Center, CH 7260 Davos Dorf, Switzerland e. Mail: cfrohlich@pmodwrc. ch; http: //www. pmodwrc. ch Report on Preliminary Results from the TSI Workshop at NIST, 18 -20 July 2005 28. 10. 2021 New. Rad 2005, Davos 17 -19 October 2005 1

Participation in the TSI WS Barnes, Robert A. , SAIC, Beltsville, Lawrence, MD George M. , LASP, Boulder, CO Bloom, Hal, NIST, Gaithersburg, Lee III, MD Robert Benjamin, Hampton Va Bosworth, John M. , Swales. Litorja, Aerospace, Maritoni, Beltsville, NIST, MD. Gaithersburg, MD Butler, James J. , EOS NASA/GSFC, Lorentz, Steven Greenbelt, R. , Standards MD. and Technology, Inc. , Ijamsville, MD. Cahalan, Robert F. , NASA/GSFC, Greenbelt, MD Lykke, Brussels, Keith, NIST, Gaithersburg, MD Crommelynck, Dominique, IRMB, Belgium Morrill, MD Jeff, Naval Research Laboratory, Washington, DC Datla, Raju, NIST, Gaithersburg, Ohno, Belgium Yoshi, NIST, Gaithersburg, MD Dewitte, Steven, IRMB, Brussels, Pankratz, Washington, Christopher DC K. , LASP, Boulder, CO Floyd, Linton, Interferometrics/NRL, Pap, Judit, Fowler, Joel, NIST, Gaithersburg, MD. NASA’s GSFC, Greenbelt, MD Parr, Albert, NIST, Gaithersburg, MD UK Fox, Nigel, National Physical Laboratory, Teddington, Middlesex Rabin, Douglas M. , NASA’s GSFC, Greenbelt, MD. Fraser, Jerry, NIST, Gaithersburg, MD. Rager, Amy, Catholic University of America, Clarksville, MD Fröhlich, Claus, PMOD/WRC, Davos Dorf, Switzerland Rice, Joe, NIST, Gaithersburg, MD Helizon, Roger S. , JPL, Pasadena, CA. Rottman, Heuermann, Karl, LASP, Boulder, COGary, LASP, Boulder, CO. Shirley, Eric, MD NIST, Gaithersburg, MD Johnson, B. Carol, NIST, Gaithersburg, Sparn, Thomas. MD P. , LASP, Boulder, CO Jordan, Stuart, NASA’s GSFC, Greenbelt, Willson, Richard C. , Columbia. MD University, Coronado, CA. Kirk, Megan, Catholic University of America, Clarksville, Kopp, Greg, LASP, Boulder, Yoon, CO Howard, NIST, Gaithersburg, MD 28. 10. 2021 New. Rad 2005, Davos 17 -19 October 2005 2

Presentations at the TSI Workshop Monday, July 18 Tuesday, July 19 Satellite Instrument TSI Measurement Uncertainty: Chair G. Kopp Wednesday, July 20 Welcome and Meeting Charge J. Butler Satellite and Ground-based TSI Instrument Comparison: Chair: R. Willson Session 1 Goals G. Kopp VIRGO/DIARAD flight performance and. Comparison degradationand S. Characterization Dewitte Laboratory-based ACRIM I, II & III R. Willson VIRGO/PMO 6 -V flight performance degradation C. Fröhlich Chair: and J. Rice ACRIM R. Helizon ACRIM 1, 2, & 3 flight comparisons, performance, degradation Aperture Area Comparisonand Results C. Johnson TIM on SORCE G. Kopp R. Willson PMO 6 V on VIRGO/So. HO C. Fröhlich Diffraction Effects E. Shirley SORCE/TIM flight comparisons, performance, and. Radiometry degradation S. G. Lorentz Kopp of Cryogenic The DIARAD type instruments, principles Applications and error estimates Shuttle TSI flight comparisons, performance, and degradation S. Dewitte NIST Cryogenic Radiometer Intercomparison J. Rice D. Crommelynck & S. Dewitte ERBS/ERBE flight performance degradation R. Lee Lab Solarand Irradiance Scale Comparisons C. Fröhlich ERBE on ERBS R. Lee Discussion and summary of flight performance and degradation Discussion and Session wrap-up Disscussion and Session wrap-up G. Kopp JPL/TMO ground comparisons and significance for flight TSI results R. Willson Workshop closing comments J. Butler Discussion and summary of ground comparisons 28. 10. 2021 New. Rad 2005, Davos 17 -19 October 2005 3

Why was the Work Shop Organized Until the launch of SORCE and the operation of TIM we all thought that we have improved the uncertainty from the early time of HF and ACRIM I. The composite shows how small the variation is compared to the spread of the data. It shows also that the precision is much higher than the absolute accuracy which allows to construct such composite with overlapping measurements. 28. 10. 2021 New. Rad 2005, Davos 17 -19 October 2005 4

Operation of Room-Temperature Radiometers Operation § § § The classical electrically calibrated radiometers are operated in either passive or active mode. Both modes have a measurement and reference phase which is determined by an open or closed shutter. In the passive mode the electrical heater is switched on when the shutter closes and its power is adjusted to produce about the same heat flux as the radiation seen during the measurement phase. In the active mode control electronics maintains the heat flux constant during both phases which are now alternated in a regular pace (e. g. 1 -min open and 1 -min closed). With the open and closed electrical power P dissipated in the cavity, the irradiance S equals with CRad the radiometer constant For an ideal radiometer CRad= 1/A with A the area of the precision aperture in front of the cavity. Deviations from the ideal behaviour are accounted for replacing 1/A by Ccorr/A with Ccorr determined for each individual radiometer by its characterization. 28. 10. 2021 New. Rad 2005, Davos 17 -19 October 2005 5

Different Types of Room-Temperature Radiometers Characterization The determination of the deviations from ideal behaviour is called characterization. The following effects are determined by independent experiments: Non-equivalence between electrical and radiative heating determined from air-to-vacuum ratio, reflectivity of the cavity, scattered light and lead heating, and by numerical calculations: Diffraction 28. 10. 2021 New. Rad 2005, Davos 17 -19 October 2005 6

VIRGO Radiometers Below is a block diagram of PMO 6 V (left) and a picture of DIARAD and PMO 6 V of VIRGO 28. 10. 2021 New. Rad 2005, Davos 17 -19 October 2005 7

Characterization of Room-Temperature Radiometers Results of a typical characterization As an example the results for the characterization of PMO 6 type radiometers is shown from Brusa & Fröhlich (1986). 28. 10. 2021 New. Rad 2005, Davos 17 -19 October 2005 8

Aperture Measurements Area of the Precision Aperture An important part of the characterization is the determination of the area of the precision aperture. In order to keep the radiometer small with a reasonable time constant the aperture is nominally 5 mm or 8 mm in diameter. The land of the aperture has to be as small as possible in order to prevent lightpipe effects due to the finite angular extent of the Sun and during slightly off-pointed conditions. Typically it is of the order of a couple 10 µm. This does not allow to measure the diameter by physically touching. Thus optical measuring methods have to be applied which are in general less accurate. Results from the presentation of C. Johnson 28. 10. 2021 New. Rad 2005, Davos 17 -19 October 2005 9

Effect of Diffraction has to be included as a correction and Eric Shirley from NIST did the calculations for all radiometers involved 28. 10. 2021 New. Rad 2005, Davos 17 -19 October 2005 10

Uncertainty PMO 6 -9 and 11 28. 10. 2021 New. Rad 2005, Davos 17 -19 October 2005 11

World Radiometric Reference and SI § From a thorough discussion of the results of comparison of many ECRs with the sun as source at PMOD/WRC the Word Radiometric Reference has been defined and adopted by WMO in 1975 and a group of radiometers identified to materialize it, the World Standard Group. The WWR is what is called in metrology a conventional reference as it has a much higher repeatability than its SI uncertainty of 0. 3%. 28. 10. 2021 New. Rad 2005, Davos 17 -19 October 2005 12

Comparison of WRR and SI, 1 of 2 Comparison with cryogenic radiometers via trap detectors at NPL have shown that the WRR is within ± 0. 02% of the SI scale (Romero et al. , 1991, 1996), which is fortuitous, but practical. These comparison are made with under-filled apertures and thus need some care about the performance of the under-filled radiometer relative to the one as measuring irradiance. Recently, such comparison were repeated with both rocket radiometers at NPL and at the Swiss Metrological Institute, METAS. The result, however, is still preliminary (see Poster 16). So we will only use the results of the 1991/96 comparison only. However, these need some corrections of the originally published values. The diffraction argument was incorrect and the new aperture areas as re-measured by NIST shall be used. Moreover, we need to apply a correction for the difference of the non-equivalence as determined from the air-to-vacuum ratio under-filled and normal. Then the results for 1991 and 1996 become 1. 000557 0. 001326 and 0. 999757 0. 000948, respectively, with a weighted average and 1 uncertainty of 1. 000091 0. 0. 000864 28. 10. 2021 New. Rad 2005, Davos 17 -19 October 2005 13

Comparison of WRR and SI, 2 of 2 Due to the small land of the precision apertures, the innermost part has a very short thermal time-constant and starts heating up when the shutter is opened. Thus the cavity receives some extra IR radiation and in air also some energy through air conduction. Although it is a rather small effect for the PMO 6 type radiometers we need to correct for it. We estimate the effect from preliminary results of model calculations which indicate that the IR influence is of the same order as the one due to the exchange in air. So the IR effect is proportional to the air-to-vacuum ratio difference of 0 820 ppm for 6 -9 and 545 770 ppm for 6 -11. Thus, a value of about 200~ppm is assumed and the final value of the WRR/SI becomes 0. 999891 with a 95% uncertainty of 0. 16%. 28. 10. 2021 New. Rad 2005, Davos 17 -19 October 2005 14

Corrections for PMO 6 V Values § § 28. 10. 2021 The radiometric constants have to be changed due to a new evaluation of the air-to-vacuum ratio and the tracing to WRR and SI. The level-1 ratio A/B is now in space 633 ppm higher than on ground. This difference can be easily explained by the uncertainty of CNE, which is for each radiometer of the order 1000 ppm (95% uncertainty). The new calibration factors change the values by 1314 ppm and 1254 ppm for A and B, respectively. The difference is mainly due to the inclusion of the diffraction correction in the radiometric constants for the comparison with the WRR. The remaining 134 and 74 ppm are due to changes of the re-evaluated air-tovaccum ratio. Finally the WRR/SI conversion has to be added and the presently adopted PMO 6 V values are increased by 1363 ppm to yield SI values New. Rad 2005, Davos 17 -19 October 2005 15

Uncertainty of PMO 6 V WRR/SI 28. 10. 2021 New. Rad 2005, Davos 17 -19 October 2005 16

Final values of VIRGO TSI at the Beginning of the Mission 1 of 2 DIARAD is fully characterized and represents its own independent radiometric scale. However, some corrections due to time constant effects and the area have been determined since the first evaluation; they amount to an overall correction of + 447 ppm. With these corrections the final ‘first light’ data of VIRGO are very consistent (within about 300 pmm), although their tracing to SI is completely independent. 28. 10. 2021 New. Rad 2005, Davos 17 -19 October 2005 17

Final values of VIRGO TSI at the Beginning of the Mission 2 of 2 If we compare this result with the internal consistency of the IRMB radiometry it may be fortuitous. These comparisons may also give a hint that the uncertainty estimate of the IRMB may be too optimistic. On the other hand, the general picture indicates that the tracing to Si via WRR is within the stated 95% uncertainties of 0. 3%. 28. 10. 2021 New. Rad 2005, Davos 17 -19 October 2005 18

Uncertainties, Correction Factors and ‘First Light’ Comparison of the Radiometers in Space 28. 10. 2021 New. Rad 2005, Davos 17 -19 October 2005 19

Comparison of corrected TSI 28. 10. 2021 New. Rad 2005, Davos 17 -19 October 2005 20

What’s next ? § § § 28. 10. 2021 It is rather obvious that power comparisons of representative spares against calibrated traps would help to locate the reason for the still unresolved difference. In contrast to the WRR/SI comparison at NPL they will be done with the radiometers in vacuum. NIST is prepared to conduct them in early 2006. As the aperture area is no longer an issue, power comparisons are sufficient. However, the implications of under-filling the apertures should be studied in detail by each radiometer involved in the comparisons. There also plans to conduct comparisons with the Sun as source at Table Mountain. We hopefully will know more in summer 2006. New. Rad 2005, Davos 17 -19 October 2005 21

Conclusions § § § 28. 10. 2021 The differences are still not resolved and the reason is unknown. It does not seem to be due to aperture areas, as it was most likely during the early measurements of TSI from space. The good agreement between PMO 6 V and DIARAD-L indicates that the WRR is most probably close to the SI as indicated by the WRR/SI comparisons. There may still be a problem with the WRR/SI comparison: the WRR factor of PMO 6 -9 and 11 are of the order of 1. 004258 and 1. 002751, indicating that the characterized PMO 6 radiometers are about 0. 35% below the WRR (similar to CROM-2 L). Because the WRR/SI comparison is only based on the PM 6 -9 and 11 radiometers, new tests can always be performed as e. g. the area of the precision aperture could be corrected. With radiometers launched to space this is obviously no longer possible and we have to rely on the similarity of the spares. New. Rad 2005, Davos 17 -19 October 2005 22

You can find this presentation at ftp. pmodwrc. ch/pub/claus/new. Rad 2005/CF_New. Rad 2005. ppt End 28. 10. 2021 New. Rad 2005, Davos 17 -19 October 2005 23

Air-to-Vacuum Ratio (Non. Equivalence) The difference between the under -filled and illuminated aperture (8. 5 mm) is 0 820 ppm for 6 -9 and 545 770 ppm for 6 -11. 28. 10. 2021 New. Rad 2005, Davos 17 -19 October 2005 24

Aperture heating 28. 10. 2021 New. Rad 2005, Davos 17 -19 October 2005 25

PMO 6 V on VIRGO § § § The main difference between the rocket PMO 6 -9, 10 and 11 and the space PMO 6 is the heat sink. The former is made out of copper and the latter from aluminum. This seems to increases the aperture heating effect. The electrical power is determined with RI 2 instead of the normal VI with R determined during each closed phase Due to the failure of the shutter operation we replaced the shutters by the covers with the following open/closed periods: • • • 28. 10. 2021 16 -Jan-1996 shutter operation of A until 4 -Feb-1996 22 -Feb-1996 change to PMO 6 V-A 8 h open and 21 min closed; PMO 6 -B 8 H closed and during 39 min centered around open of A 6 -Jul-1996 change to PMO 6 -A 8 h open and 6 min closed; PMO 6 -B once per week during 33 min open, still centered around open of A New. Rad 2005, Davos 17 -19 October 2005 26

PMO 6 V on VIRGO: Electrical measurements 28. 10. 2021 New. Rad 2005, Davos 17 -19 October 2005 27

PMO 6 V on VIRGO level-1 Data A major problem are changes of sensitivity of the radiometers due to the exposure to the strong UV radiation of the Sun and the exposure to the space environment. In the present context we are only interested in the early increase. Early increase of PMO 6 V-A 28. 10. 2021 New. Rad 2005, Davos 17 -19 October 2005 28

PMO 6 V on VIRGO: Non-equivalence 28. 10. 2021 New. Rad 2005, Davos 17 -19 October 2005 29

PMO 6 V on VIRGO: Radiometric constants 28. 10. 2021 New. Rad 2005, Davos 17 -19 October 2005 30

Influence of new operation 28. 10. 2021 New. Rad 2005, Davos 17 -19 October 2005 31

Changes of PMO 6 V final values The Feb/March 1996 irradiance become for PMO 6 V-A level-1. 8 1366. 785 and for PMO 6 V-B 1366. 782 with a mean of 1366. 784 Wm-2. These new radiometric constants need to be referred to SI by the WRR/SI factor. To convert the WRR value into SI we devide the values by 0. 9998322 found from WRR/SI from 3 independent comparisons. This yields 1367. 013 Wm-2 which is 1408 ppm higher than the present PMO 6 V level 1. 8 data. 28. 10. 2021 New. Rad 2005, Davos 17 -19 October 2005 32

Uncertainty of Ground-based Calibrations 28. 10. 2021 New. Rad 2005, Davos 17 -19 October 2005 33