How Does the Sun’s Spectrum Vary? Modeling Solar Cycle Irradiance Variations Judith Lean, Karl Battams Space Science Division, Naval Research Laboratory, Washington DC Odele Coddington, Gary Rottman, Peter Pilewskie LASP, University of Colorado, Boulder CO Solar Rotation – days to months, multiple "realizations" - relatively well specified - issues remain about spectral dependencies – UV vs IR - Marchenko, DeLand and Lean, Space Weather and Climate, 2017 Solar Cycle – years to decades - challenged by instrument sensitivity drifts - disagreement among observations and models Primary Focus of 2nd Year Effort Long Term – multiple decades - speculative; depends on constraining & understanding solar cycle variations Funded by NASA SIST Meeting, Greenbelt, MD, 11-12 May 2017 validation TIM TSI Observations vs NRLTSI2 (NOAA CDR) model training ΔTSI(t) = ΔTSIfaculae(t) +ΔTSIspot(t) NRLTSI2 model systematically overestimates TIM TSI at solar minimum by ∼0.05 W m-2 (6% of solar cycle variation) - spots? - faculae? - TIM? - something else? SOLSTICE Lyman α Observations vs NRLSSI2 ΔFLy𝜶𝜶(t) = ΔTSIfaculae(t) = a + b× ΔMg(t) validation NRLSSI2 Lyman 𝜶𝜶 irradiance is linear function of Mg index, proxy for faculae model formulation NRLSSI2 model (faculae) underestimates SOLSTICE Ly𝜶𝜶 cycle 23 decline - faculae? - SOLSTICE? - something else? NRLSSI2 model systematically overestimates SOLSTICE Ly𝜶𝜶 at solar minimum by ∼0.1 mW m-2 (5% of solar cycle variation)…. when normalized to cycle 24 - consistent with TIM vs NRLTSI2 TIM Bolometric Faculae vs. SOLSTICE Lyman α ΔTSIfaculae(t) = ΔTSI(t) -ΔTSIspot(t) Residual trend of 0.15 W m-2 per decade is 7.5% of solar cycle amplitude - SOLSTICE Lyman α overestimates cycle 23 decline relative to TIM bolometric faculae - good agreement at solar minimum High correlation of daily TIM bolometric faculae and SOLSTICE Lyman α irradiance TIM Bolometric Faculae: Insensitive to Sunspot Darkening Index Formulation Bolometric facular variability: ΔTSIfaculae(t) = ΔTSI(t) - ΔTSIspot(t) ΔTSIspot(t) ∝ΣAsCs μ(3μ+2)/2 all observatories As=sunspot area four main Cs=sunspot contrast observatories area independent area dependent μ=cos(lat)cos(long) location at time of observation adjusted to central meridian sunspot index CDR –all observations, area-independent contrast, observed location four main observations, area-independent contrast, adjusted to central meridian Dependence of TIM Bolometric Faculae and SOLSTICE Lyman α on Mg Index For both TIM bolometric faculae and SOLSTICE Lyman α irradiance, quadratic parameterization of Mg index is superior to linear parameterization… improved facular parameterization might be ΔFLy𝜶𝜶(t) = a + b×ΔMg(t) + c×[ΔMg(t)]e New Model of TSI Variability: NRLTSI2e ΔTSI(t) = ΔTSIfaculae(t) + ΔTSIspot(t) e=1.2 = a + b×Mg(t) + c×Mg(t)e + d×Ps(t) Correlation with TIM improves from 0.960 to 0.969 Standard deviation of residuals decreases from 0.121 to 0.107 W m-2 Residual trend remains better than TIM's 10 ppm per year repeatability New Model of SSI Variability: NRLSSI2e F(λ,t) = Fquiet(λ) + ΔFfaculae(λ,t) + ΔFspot(λ,t) OR ΔFfaculae(λ,t) ∝ ΔTSIfaculae(t) = 296.6×Mg(t) + 377.5×Mg(t)1.2 ΔFfaculae(λ,t) ∝ ΔLyαfaculae(t) = 345.2×Mg(t) + 376.9×Mg(t)1.2 Correlation of Lyα models using two different ΔF faculae is 0.9997 Correlation with SOLSTICE improves from 0.987 to 0.989 Model constructed using direct (not detrended) SOLSTICE Lyα observations Standard deviation of residuals decreases from 0.119 to 0.106 mW m-2 ….slope of residuals is 0.15% per year Lyα Comparison: NRLSSI2e vs Observations will NRLSSI2e agree better with …. - new SIST Lyα composite? - SOLID composite? SME UARS SORCE slope of residuals is 0.3% per year will new SIST Mg composite improve model agreement with observations? Solar Cycle Irradiance Variations: 240-241 nm Direct Observations Detrended Observations Use ∆Mg -cycle to estimate ∆UV-cycle Determine ∆UV for ∆Mg - rotation Testing the Scaling of Rotational Modulation to Solar Cycle Modulation ΔF Ly𝜶𝜶 =a+ b×Mg(t) ΔFLy𝜶𝜶 =a+ b×Mg(t) + c×Mg(t)1.2 How Does the Sun’s Spectrum Vary? SUMMARY, Year 2 Modelling Solar Cycle Irradiance Variations - demonstrated the mutual consistency of the SORCE TIM TSI and SOLSTICE Lyman a SSI observations over the 11-year cycle - identified the cause of differences between NRL's CDR models and SORCE TSI and Lyman a observations - constructed an improved facular index by adding an exponential component of the Mg index to model formulation - evidence for long-term trend in SOLSTICE Lyman a <0.15% per year - determined new time-dependent coefficients for scaling facular index from rotation to solar cycle variations, for modeling SSI at longer wavelengths where observations are less stable Solar Rotation -Marchenko, DeLand and Lean, Space Weather and Climate, 2017 Long Term - new white-light coronal irradiance index suggests ACRIM TSI inter-minim trend from 1996 to 2008 is too big How Does the Sun’s Spectrum Vary? YEAR 3 WORK - Construct a new version NRLSSI2e spectral irradiance variability model (and uncertainties), over multiple solar cycles, and compare with observations (revised from SIST activities). - Compare new SME database to measurements of spectral irradiance from other instruments for concurrent time, and with new model - Continue to validate and explore improvements for sunspot and facular indices and rotationto-cycle scaling - Prepare paper(s) describing results of solar cycle change in comparison with existing models of solar spectral irradiance variability, and independent solar irradiance observations. - Revise and resubmit paper about analysis off LASCO solar irradiance index comparisons with direct observations In collaboration with other SIST members:. - Incorporate new total solar irradiance composite to additionally constrain solar cycle spectrum changes - Analyze and compare facular component of new total solar irradiance composite - Compare new Lyman α composite with sunspot-corrected new total solar irradiance composite