Solar Ultraviolet Irradiance Variations over Four Solar Cycles Tom Woods Laboratory for Atmospheric and Space Physics University of Colorado email: tom.woods@lasp.colorado.edu SSI Data Sets (http://lasp.colorado.edu/lisird) SME: G. Rottman UARS SOLSTICE: G. Rottman & T. Woods UARS SUSIM: L. Floyd TIMED SEE: T. Woods & F. Eparvier SORCE SOLSTICE: M. Snow & W. McClintock SORCE SIM: J. Harder & J. Fontenla Composite SSI (SBUV): M. DeLand SSI Models NRLSSI: J. Lean SATIRE: W. Ball SRPM: J. Fontenla W&R: T. Woods & G. Rottman
Aside on lower TSI Level LASP s TSI Radiometric Facility (TRF) has been key in providing new TSI instrument calibrations for Glory TIM, Picard PREMOS, SOHO PMOD, and ACRIM New result is lower TSI level of ~1361 W/m 2 Kopp & Lean (GRL, 2011)
Outline for Solar Spectral Variations Examples of solar cycle variability Composite H I Lyman-α (121.6 nm) Harder et al. (2009) SORCE SIM Result Solar cycle variability versus degradation Examining solar variability across solar cycle minimum (SCM) SME: SCM 21-22 UARS: SCM 22-23 SORCE and TIMED: SCM 23-24 Composite solar cycle variability Comparison to SBUV observations Comparison to models NRLSSI, SRPM, SATIRE, W&R
Composite Solar UV Variability Example: H I Lyman-α by combining 5 different measurement sets and filling gaps with empirical models using proxies (e.g. F10.7) M. DeLand also has made composite solar UV (200-400 nm) from 1978 to 2005. These composites can be downloaded from http://lasp.colorado.edu/lisird/
Example solar cycle (SC) variability UV UV varies the most when variability is given as a ratio (percentage) However Variability given in energy units is more appropriate for climate studies Note that red lines are variations that are out of phase with the solar cycle. Solar Cycle (SC) variability from the NRLSSI model (J. Lean)
new SSI variability results from SORCE In addition to the infrared, some visible wavelengths are out of phase with the solar cycle Also, there is more UV variability than expected Are the SIM and model differences possibly related to unresolved instrument trends? Checking these results is challenging Other validation techniques are needed than direct comparisons Figures from J. Harder et al., GRL, 2009 Out of Phase Wavelengths
Comparison of Many Recent Studies Fig. 2 from TOSCA paper (Ermolli et al. 2012) TOSCA May 2012 Workshop : Recent variability of the solar spectral irradiance and its impact on climate modelling Harder et al. 2009 Wavelength (nm) Next slides discuss this SORCE reanalysis which is really analysis of SC variability measurements from 1984-2012
Understanding degradation is critical for understanding solar variations Measurement = SIM uncorrected data at 280 nm
Technique: pick days of similar solar activity levels equal-distance from Min. SC 21 22 23 24 Proxy composite: <TSI> <27 days> SSN (daily) <SSN> <Mg C/W> <Lyman-α> F10.7 <F10.7> Range: 0-700 as each scaled to 100 for its SC variability range Reference Level is somewhat arbitrary Picked 2004/255
Assumption: average of variability before and after minimum = real solar variability Average Variability 1.5% (corrected) Versus Var. = Max/Min - 1 1.4% 1.6% 1.9% (raw data) 8.0% -4.1% Technique Uncertainty is ~20%
This technique can be tested with missions that have measured over cycle minimum SC 21 22 23 24 SME UARS SORCE ASIDE: Proxy composite has lower minimum in 2008, so variability for SC 23-24 can be relatively larger for this technique (e.g. PMOD ΔTSI is x 2 larger) Proxy composite: <TSI> <27 days> SSN (daily) <SSN> <Mg C/W> <Lyman-α> F10.7 <F10.7> Range: 0-700 as each scaled to 100 for its SC variability range Reference Level is somewhat arbitrary Picked 2004/255
PMOD TSI suggests same levels are reasonably selected. 3-σ variation for detrended 27-day smoothed TSI = 313 ppm
Solar Images for same levels
SSI data sets Fig. 3 in TOSCA paper (Ermolli et al. 2012) Solar Cycle (SC) 21/22 22/23 23/24 SC 21-24 Composite (Woods SORCE reanalysis) includes SME 1984-1988 UARS SOLSTICE & SUSIM 1994-1998 TIMED SEE 2004-2012 SORCE SOLSTICE & SIM 2004-2012
SME over solar cycle min 21-22 Ideal Comparison Best Wavelengths 140-210 nm
UARS SOLSTICE SC Min 22-23 Best Wavelengths 120-170 nm
UARS SUSIM SC Min 22-23 Best Wavelengths 140-260 nm
TIMED SEE SC Min 23-24 Best Wavelengths 0-140 nm
SORCE L3 SC Min 23-24 SORCE L3 SOLSTICE: 115-308 nm SIM: 308-1600 nm Best Wavelengths 115-270 nm & 310-400 nm
SORCE SIM L2 SC Min 23-24 degradation trend fit to same levels first Best Wavelengths 260-400 nm
Combine the best of each to make Composite Solar Cycle Variability Note that full solar cycle variation is about a factor of 3 more
Combine the best of each to make Composite Solar Cycle Variability Note that full solar cycle variation is about a factor of 3 more
ΔSSI / ΔTSI from this composite SSI 1-σ uncertainty is about ±30% Wavelength (nm) 0-200 nm is 1.2% of ΔTSI 0-300 nm is 28% 0-400 nm is 116% So the 300-400 nm range is most important for this comparison to TSI
DeLand SBUV-UARS SSI Comparison to this new Composite Measurement Best Wavelengths for SBUV 140-200 nm
NRLSSI model (J. Lean) Comparison to Composite Measurement
SATIRE model (W. Ball) Comparison to Composite Measurement
SRPM model (J. Fontenla) Comparison to Composite Measurement
Woods & Rottman (2002) model Comparison to Composite Measurement
All Models compared to Composite 0-220 nm
All Models compared to Composite 180-400 nm
Conclusions Observations over a solar cycle minimum permit independent check on degradation trend (at ~20% of the variability) Solar ultraviolet variability observations and models agree best at wavelengths shorter than 210 nm No clear consistency found between observations or models for wavelengths longer than 260 nm Additional validation / analyses needed for the 200-400 nm range of past missions. Additional (new) data sets are also needed.
Future SSI Measurements Can new instruments have little or no degradation for the UV-Vis-NIR? NOAA/NASA TSIS SIM 3 channels: daily, monthly, annual Improved lower-noise version of SORCE SIM Féry prism spectrometer covering the full wavelength range from the UV to IR using only one optical element for spectral dispersion and image quality
NUV-Vis-NIR Comparisons preliminary results
SORCE L3 SC minimum technique result ΔTSI 0.032%
SORCE SIM L2 degradation trend fit to same levels first ΔTSI 0.032%
All Models compared to Composite 400-1600 nm ΔSSI / ΔTSI 0-1600 nm 96% 130% 104% N/A 84% ΔTSI 0.032% Note that SIM 0-1600 nm SSI / TSI = 88% (expected for values above)
Backup Slides
List of days of same level Solar Cycle Minimum 1986/195 1996/140 2008/280