SOLAR SPECTRAL IRRADIANCE VARIABILITY OF saj.matf.bg.ac.rs/194/pdf/071-086.pdf · SOLAR SPECTRAL IRRADIANCE…

Download SOLAR SPECTRAL IRRADIANCE VARIABILITY OF saj.matf.bg.ac.rs/194/pdf/071-086.pdf · SOLAR SPECTRAL IRRADIANCE…

Post on 10-Aug-2018

212 views

Category:

Documents

0 download

TRANSCRIPT

<ul><li><p>Serb. Astron. J. } 194 (2017), 71 - 86 UDC 523.945 + 523.98DOI: https://doi.org/10.2298/SAJ160126001G Professional paper</p><p>SOLAR SPECTRAL IRRADIANCE VARIABILITYOF SOME CHROMOSPHERIC EMISSION LINES</p><p>THROUGH THE SOLAR ACTIVITY CYCLES 21-23</p><p>U. D. Goker1,2, M. Sh. Gigolashvili3 and N. Kapanadze3</p><p>1Physics Department, Bogazici University, Bebek 34342, Istanbul, Turkey2Istanbul Gelisim University, Faculty of Economics, Administrative andSocial Sciences, Department of Aviation Management, Cihangir Quarter,</p><p>Sehit Jandarma Komando Er Hakan Oner Street 1, Avclar/Istanbul, TurkeyEmail: udgoker@gelisim.edu.tr</p><p>3E. Kharadze Abastumani Astrophysical Observatory at Ilia State University,Kakutsa, Cholokashvili Ave 3/5, Tbilisi, 0162, Georgia</p><p>Email: natela.kapanadze@iliauni.edu.ge</p><p>(Received: January 26, 2016; Accepted: January 16, 2017)</p><p>SUMMARY: A study of variations of solar spectral irradiance (SSI) in the wave-length ranges 121.5 nm-300.5 nm for the period 1981-2009 is presented. We usedvarious data for ultraviolet (UV) spectral lines and international sunspot number(ISSN) from interactive data centers such as SME (NSSDC), UARS (GDAAC),SORCE (LISIRD) and SIDC, respectively. We reduced these data by using theMATLAB software package. In this respect, we revealed negative correlations ofintensities of UV (289.5 nm-300.5 nm) spectral lines originating in the solar chro-mosphere with the ISSN index during the unusually prolonged minimum betweenthe solar activity cycles (SACs) 23 and 24. We also compared our results with thevariations of solar activity indices obtained by the ground-based telescopes. There-fore, we found that plage regions decrease while facular areas are increasing in SAC23. However, the decrease in plage regions is seen in small sunspot groups (SGs),contrary to this, these regions in large SGs are comparable to previous SACs oreven larger as is also seen in facular areas. Nevertheless, negative correlations be-tween ISSN and SSI data indicate that these variations are in close connection withthe classes of sunspots/SGs, faculae and plage regions. Finally, we applied the timeseries analysis of spectral lines corresponding to the wavelengths 121.5 nm-300.5nm and made comparisons with the ISSN data. We found an unexpected increasein the 298.5 nm line for the Fe II ion. The variability of Fe II ion 298.5 nm line is inclose connection with the facular areas and plage regions, and the sizes of these solarsurface indices play an important role for the SSI variability, as well. So, we com-pared the connection between the sizes of faculae and plage regions, sunspots/SGs,chemical elements and SSI variability. Our future work will be the theoretical studyof this connection and developing of a corresponding model.</p><p>Key words. Methods: data analysis Sun: chromosphere Sun: UV radiation</p><p>This paper is dedicated to the memory of Prof. Marina Sh. Gigolashvili.</p><p>c 2017 The Author(s). Published by Astronomical Observatory of Belgrade and Faculty of Mathematics, University ofBelgrade. This open access article is distributed under CC BY-NC-ND 4.0 International licence.</p><p>71</p></li><li><p>U. D. GOKER et al.</p><p>1. INTRODUCTION</p><p>The variability of solar spectral irradiance(SSI) in ultraviolet (UV) spectral range was signifi-cantly stronger in the last solar activity cycle (SAC)23 than in the previous SACs 21 and 22. However,the cycle 23 did not behave as predicted: it was amagnetically weaker cycle and complex in terms ofmagnetic morphology (Gibson et al. 2011), it wasweak in sunspot number and facular activity (deToma et al. 2004), Mg II index and 10.7 cm solarradio flux (F10.7) than the two previous cycles 21and 22 (Fisk and Zhao 2009). On the other hand,as has been argued in Krivova et al. (2003), theSSI variations in cycle 23 could not be explained interms of surface magnetic features alone. It impliesanother or additional mechanisms such as sunspots,faculae, plages and chromospheric network for irra-diance change.</p><p>We investigate an unexpected increase in SSIintensity for some spectral lines in the SAC 23from our data analysis and correlate this increasewith faculae and plage regions, and sunspot/sunspotgroup numbers and classes. In previous studies, theanomaly in SAC 23 was correlated with the inter-national sunspot number (ISSN) observations alonerather than taking into account different sunspotsand sunspot group classes and their structures, sep-arately. The very specific and important studies ofLefevre and Clette (2011), Kilcik et al. (2011) andKilcik et al. (2014) showed us the importance of thenumber and the class of small and large sunspotsand/or sunspot groups (SGs) on the solar disc andtheir effects on the SACs.</p><p>In addition to this, helio-seismic data havealso showed a surprising anti-correlation betweenfrequencies and activity proxies during the mini-mum between SACs 23 and 24 (Tripathy et al.2010). The heavy element abundances affect thehelio-seismology by causing changes in the convec-tion zone structure, and changes in the convectionzone will affect the radiative opacities, the boundarybetween the radiative and convective zones and theequation of state (Basu and Antia 2004). The carbongroup elements (e.g. carbon, silicon, germanium, tinand lead) and the bound-bound absorption of themetallic lines below 400 nm are the most importantopacity sources (Fontenla et al. 1999).</p><p>From all these different points of view, theaim of this paper is the comparison of some selectedSSI intensities for SACs 21-23 in the UV spectralrange with the possible variations of chemical ele-ment ions corresponding to these wavelength ranges.We mainly concentrate our attention on this vari-</p><p>ation, in addition to this, we apply the variationsof sunspots/SGs, plage regions, faculae and Ca IIK-flux on the anomalies of the SSI variability be-tween the spectral ranges 121.5 nm-300.5 nm andthe changes in related chromospheric ions betweenthe years 1981-2010.</p><p>The organization of the paper is as follows. InSection 2, we give the data description, clarify thedata analysis technique and present the data analy-sis results for SSI, ISSN and spectral lines in the firstpart; we compare our results with sunspots/SGs,plage regions, faculae and Ca II K-flux which weretaken by the ground-based telescopes in the secondpart. In Section 3, discussions about the data analy-sis results are presented and in Section 4, conclusionsare given.</p><p>2. DATA DESCRIPTION, DATAANALYSIS AND RESULTS</p><p>2.1. Data description and data analysistechnique</p><p>The peculiarities of SAC 23 in the UV spec-tral range caused us to investigate the SSI variationsfor wavelengths 121.5 nm-300.5 nm to reveal possi-ble special features of irradiance variations throughSAC 23. Based on consideration of the abnormal-ity of SAC 23, we decided to find out how well theirradiance variability of different spectral bands co-incides with the ISSN for different SACs 21-23 andwe estimated the existing correlations quantitatively.</p><p>Firstly, we investigated the SSI intensitiesin the range of 121.5 nm-300.5 nm for SACs 21,22 and 23 obtained by Solar Mesosphere Explorer(SME) for the years 1981-1990 (the available wave-length range 115.5 nm-302.5 nm), Upper Atmo-sphere Research Satellite (UARS) for the years1992-2001 (the available wavelength range 119.5 nm-425.5 nm) and The Solar Radiation and ClimateExperiment (SORCE) for the years 2003-2009 (theavailable wavelength range 116.5 nm-1598.95 nm)experiments. Time series of SSI have been ex-tracted from data centers of National Space ScienceData Center (NSSDC), Goddard Distributed Ac-tive Archive Center (GDAAC) and LASP Interac-tive Solar Irradiance Data Center (LISIRD) thatserved the archives of NASA space science missiondata. The data contain hourly and daily measure-ments, and they are merged into ASCII text-filesmaking data convenient. These different interactivedata centers have taken the observational data foreach 0.5 nm in the UV spectral range, so we made</p><p>http://www.jpl.nasa.gov/missions/solar-mesosphere-explorer-sme/http://uars.gsfc.nasa.gov/http://lasp.colorado.edu/home/sorce/http://nssdc.gsfc.nasa.gov/http://daac.gsfc.nasa.gov/http://lasp.colorado.edu/lisird/</p><p>72</p></li><li><p>SOLAR SPECTRAL IRRADIANCE VARIABILITY OF SOME CHROMOSPHERIC EMISSION LINES</p><p>our analysis for each 0.5 nm wavelength interval. Alist of the selected data for UV spectral range usedin the present analysis is given in Table 1. In thistable, space flight experiments and data archives areindicated as well. The data processing includes:(a) extracting and compiling of selected quantities;(b) creation of the homogeneous local data set of for-matted data;(c) statistical processing of investigated data (regres-sion analysis, correlation analysis).</p><p>The determination of SSI variability and find-ing accurate results below 400 nm is not easy. In-vestigations of de Toma et al. (2001) illustrate the</p><p>difficulty in using simple proxies and regression tech-niques to deduce physical sources of SSI variability.Dudok de Wit et al. (2009) indicate that the bestresults for SSI have been taken from experimentaldata by using statistical analysis or by computingsome special events. Besides analyzing techniques,there is another important problem with the choiceof minimum days because the day-to-day solar irradi-ance variability is small at low levels of solar activity(Floyd et al. 1998). From this respect, we had toeliminate or reduce to a minimum the numbers oferrors for these important problems to analyze theSSI variability.</p><p>Table 1. List of the selected data for UV spectral ranges used in the present analysis is given. The spaceflight experiments and data archives are indicated as well.</p><p>Selected Datafor Processing The Spaceflight Experiments and Data Archivesand Origin Region</p><p>1981-1986 1987-1990 1992-1996 1997-2001 2003-2009cycle 21-Da cycle 22-Ab cycle 22-D cycle 23-A cycle 23-D</p><p>121.5 nm (Lyman ) SME SME UARS UARS SORCEemission (NSSDC) (NSSDC) (GDAAC) (GDAAC) (LISIRD)Trans. Region200.5 nm - 300.5 nm SME SME UARS UARS SORCEemission (NSSDC) (NSSDC) (GDAAC) (GDAAC) (LISIRD)ChromosphereISSN SIDC SIDC SIDC SIDC SIDCSolar Surface</p><p>a Descendingb Ascending</p><p>Table 2. Correlation coefficient values (with standard errors) correspond to the intensity profiles of theselected spectral lines with ISSN.</p><p>Spectral Lines Correlation Coefficients (rc)1981-1986 1987-1990 1992-1996 1997-2001 2003-2009</p><p>121.5 nm 0.930.01 0.940.01 0.930.01 0.890.01 0.830.01285.5 nm 0.760.02 0.310.03 0.370.03 0.540.02 0.030.02300.5 nm 0.180.02 0.340.03 0.500.02 0.060.03 -0.740.01</p><p>Table 3. Correlation coefficient values (with standard errors) correspond to the standardized daily valuesof the selected spectral lines with ISSN.</p><p>Spectral Lines Correlation Coefficients (rc)1981-1986 1987-1990 1992-1996 1997-2001 2003-2009</p><p>121.5 nm (Lyman ) 0.930.01 0.940.01 0.930.01 0.890.01 0.830.01246.5 nm (Fe II) 0.870.01 0.900.01 0.860.01 0.850.01 0.770.01279.5 nm (Mg II) 0.760.01 0.310.03 0.750.02 0.550.02 0.030.02286.5 nm (Cr II) 0.660.02 0.280.03 -0.110.02 0.240.03 0.720.02298.5 nm (Fe II) 0.420.02 0.030.03 -0.200.02 0.150.03 -0.730.02</p><p>73</p></li><li><p>U. D. GOKER et al.</p><p>Table 4. The estimation of coefficients of linear fits (with their errors) for relationships between SSI andISSN for the SACs 21-23. The unit of slope B is given by W/m2/nm.</p><p>I.P.S.S.L. Coefficients of Linear Fits1981-1986 1987-1990 1992-1996 1997-2001 2003-2009</p><p>121.5 nm B (-3.360.07)105 (2.600.06)105 (-5.560.14)105 (5.080.08)105 (-6.020.06)105A (5.460.23)103 (4.970.26)103 (7.940.26)103 (7.450.45)103 (6.940.36)103</p><p>285.5 nm B (-5.130.03)104 (4.260.02)105 (9.620.01)105 (1.440.01)105 (-1.610.01)104A (6.930.60)104 (0.150.13)103 (0.325.39)104 (0.120.04)103 (0.360.03)104</p><p>300.5 nm B (-1.270.02)103 (2.700.01)104 (8.550.05)105 (1.040.34)105 (2.380.04)104A (8.553.75)104 (1.300.41)104 (3.090.39)104 (3.851.20)104 (1.170.09)104</p><p>S.D.V.S.S.L. Coefficients of Linear Fits1981-1986 1987-1990 1992-1996 1997-2001 2003-2009</p><p>121.5 nm B (-3.360.07)105 (2.600.06)105 (-5.560.14)105 (5.080.08)105 (-6.020.06)105A (5.460.23)103 (4.970.26)103 (7.940.26)103 (7.450.45)103 (6.940.36)103</p><p>246.5 nm B (-0.670.01)105 (0.530.01)105 (-0.790.01)105 (0.860.01)105 (-2.860.02)105A (5.380.02)103 (2.590.14)103 (4.060.18)103 (6.620.48)103 (6.350.44)103</p><p>279.5 nm B (-0.680.01)105 (0.530.01)105 (-0.800.02)105 (0.830.01)105 (-2.860.02)105A (5.370.02)103 (2.580.20)103 (4.050.18)103 (6.620.47)103 (6.340.44)103</p><p>286.5 nm B (-5.100.03)104 (9.090.02)104 (9.600.03)104 (4.350.01)104 (-1.600.01)104A (3.250.28)103 (7.292.62)104 (8.335.37)104 (3.071.26)104 (5.350.03)104</p><p>298.5 nm B (-1.270.02)103 (2.700.01)104 (8.300.05)105 (1.030.04)103 (3.400.04)104A (2.553.75)103 (1.300.41)104 (3.120.39)104 (6.351.18)104 (1.830.21)104</p><p> Intensity Profiles of the Selected Spectral Lines (Table 2) Standardized Daily Values of the Selected Spectral Lines (Table 3)</p><p>We chose the following wavelengths from theSSI database: 121.5 nm, 285.5 nm and 300.5 nmto analyze and compare the SSI variations of theselected bands with the chromospheric ions, andISSN for the wavelenghts 121.5 nm, 246.5 nm, 279.5nm, 286.5 nm and 298.5 nm. The ISSN data wereobtained from the Solar Influences Data AnalysisCenter (SIDC) and World Data Center for theSunspot Index, at the Royal Observatory of Bel-gium. The 121.5 nm Lyman- emission originatesin the transition region of the Sun while 246.5 nm,279.5 nm, 285.5 nm, 286.5 nm, 298.5 nm and 300.5nm emission components of the SSI spectral linesoriginate in the solar chromosphere. Selected ASCIIdata have been reduced using MATLAB computinglanguage. The statistical processing of the selecteddata includes the following steps:</p><p>(a) Interpolation of missing data: Lineardata interpolation was used in the case when datawere missing. We interpolated daily values of dataindices to get continuous and spectrally resolvedmeasurements of the irradiance and more obvioustrends of time series for SACs 21, 22 and 23.</p><p>(b) Primary processing of selected data(timed average, smoothed average): Since therewere several data for a given day, we used the av-erage values. The SSI data after the interpolationand timed/smoothed averaged processing are givenin Figs. 1-3. These data will be used for regressionanalysis and standardization.</p><p>(c) Regression analysis of selected datato reveal linear trends: For the linear regressionwe first calculated the estimated A and B coeffi-cients:</p><p>y = A + Bx , (1)to find the equation of the best fit line above and therelationship between selected SSI and ISSN data by:</p><p>A =n</p><p>i=1</p><p>(x2i y (xiyi)x</p><p>)</p><p>, (2)</p><p>B =n</p><p>i=1 (xiyi xiy)</p><p>, (3)</p><p>2A =2yN</p><p>(ni=1 x</p><p>2i</p><p>), (4)</p><p>2B =</p><p>(2y</p><p>), (5)</p><p>where the independent variable xi corresponds to10-day averaged values selected from the ISSN dataand the dependent variable yi corresponds to the SSIdata, N is the total number of data, i indicates thenumber of measurements, A is the deviation of se-lected data for coefficient A, B is the deviation ofselected data for coefficient B and y is the standarddeviation for the selected SSI data. and y aregiven by:</p><p> =n</p><p>i=1</p><p>(x2i xix</p><p>), (6)</p><p>http://sidc.oma.be/http://si...</p></li></ul>

Recommended

View more >