Variations in the Thermodynamic State of the Chromosphere over the Sunspot Cycle
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
Smithsonian Contributions to Astrophysics Vol. 2, No. 4. 1957 Variations in the Thermodynamic State of the Chromosphere over the Sunspot Cycle By R. G. Athay,1 D. H. Menzel,2 and F. Q. Orrall 2 Introduction sphere from observations over a narrow sector The solar corona undergoes marked syste- of the limb at a given eclipse? If we can, does matic changes in brightness and shape during this average change with time? In this paper the course of a sunspot cycle. Standard we shall attempt to answer both questions. photometric techniques enable us to detect the In order to study chromospheric changes, we variations in both the continuous and the line need reliable emission gradients and reliable spectra. The flash spectrum observed at absolute and relative intensities of both line eclipse indicates that the chromospheric spec- and continuous emission. Data from five trum also undergoes marked changes in char- eclipses are available for study, not uniformly acter (Menzel, 1931; Cillie and Menzel, 1935; precise as far as photometric standardization is Athay and Thomas, 1956). The significance of concerned. The 1932 eclipse occurred within a this apparent variation in chromospheric struc- year of sunspot minimum, the 1941 and 1952 ture is somewhat difficult to assess, in terms of eclipses two years before minimnTn, the 1945 the physical character. The available data eclipse one year after minimum, and the 1936 concerning chromospheric variability are ex- eclipse one year before maximum. The avail- tremely limited. One may reasonably question able data for the 1945 eclipse are very limited. the reliability of the indicated changes. The Thus, evidence for time-variable changes in available data refer to the chromosphere in the chromospheric structure must depend heavily equatorial regions. To our knowledge, no on the 1932 and 1936 eclipses. Fortunately, eclipse observations exist for the polar chromo- "jumping-film" spectrograms for those two sphere. Existing spicule and prominence struc- eclipses and the 1952 eclipse are at our disposal. ture must produce some variation in the char- As a result, we shall center our discussion acter of the spectrum from point to point on the around these data. limb. Thus, we must find the answer to two We have mentioned three quantities of inter- main questions, in our search for systematic est in the character of the chromospheric spec- changes in chromospheric structure. Can we trum: emission gradients, relative intensities, determine the properties of an average chromo- and absolute intensities. Any attempt to set 1 On leave of absence from High Altitude Observatory. > Harvard College Observatory and Sacramento Peak Observatory. 410217—57 8 35
36 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS the relative intensities on an absolute scale en- across the solar surface. We assume that the counters serious difficulties. Since the various brightness of this coronal image is constant emission lines may possess different gradients, with time and that the color at all heights any change in the zero point of the height scale matches that of the disk. We can determine will produce systematic differences in both rela- the absolute brightness of the coronal image tive and absolute intensities. A shift in zero as a function of distance from the center of point by only 300 km leads to apparent changes the disk, either from direct measurements made of a factor two in the ratio of intensities of he- at the eclipse or from van de Hulst's (1953) lium lines to faint metal lines. For this reason, coronal model for the polar regions. Independ- we have redefined the zero point of the height ent measurements of coronal brightness are scale for the 1932 and 1936 eclipses to be con- available for 1936. For 1932 and 1952 we must sistent with that used for 1952. rely on the coronal models. Although considerable effort has gone into The corona serves as a continuous light source the reduction of the 1936 data, no results have with an intensity variation perpendicular to the been published in readily available form. direction of dispersion. In this respect it is Hemmendinger (1939) measured line intensities equivalent to a uniform source exposed through at several points on the limb and Menzel (un- a slit of varying aperture. The intensity dis- published data, 1939) made similar measure- tribution with wavelength is accurately known ments for two points on the limb. The advent since it corresponds to that of the uneclipsed of World War II interrupted the analysis and sun. Although the intensity distribution with reduction of these data. Since we are looking distance from the center of the disk and the specifically for evidence of changes in chromo- absolute intensities are less accurately known, spheric structure, it seemed advisable to re- we may still employ them as a secondary measure some of the data for the 1936 and 1932 standard. Since light rays from both corona eclipses in order to obtain as much homogeneity and chromosphere trace essentially identical as possible in the reduction techniques. The paths through the earth's atmosphere and techniques used were consistent with those used through optical instruments, the corrections for for 1952. atmospheric extinction, instrumental absorp- We have mentioned that reliability of the tion, and film sensitivity are automatically in- photometric standardizations is a prime ques- cluded. Also, since the coronal standard and tion in all existing eclipse data. In the follow- thromospheric spectrum receive identical expo- ing section we shall review the standardizing sures and development, the determined coronal methods used at the three eclipses, and shall intensity serves as an additional check on the discuss an alternative method of standardiza- constancy of successive exposures. tion which indicates the reliability of the data. In slitless spectrograms each wavelength pro- duces a ring-shaped image of the corona. The Photometry continuum intensity at any point in the spec- Before we discuss the methods of standardiza- trogram is the sum of the overlapping im- tion at the three eclipses we shall consider how ages. Each image represents a different wave- to check the reliability of the results. One of length and different point in the coronal im- us (Athay, 1953) has previously pointed out age. At the limb ± 90° to the line of dispersion, that the coronal spectra superposed on slitless the images are displaced along a tangent to the eclipse spectrograms may be used as a second- limb. The tangential scale height in the lower ary standard light source. Here we shall only corona is about 0.8 solar radii. Thus, most of briefly summarize the method and assumptions the intensity at a point beyond the limb at used. ±90° to the line of dispersion will be built up The corona as a standard source.—On slitless from images displaced less than a solar radius. spectrograms with the dispersion set parallel to On the spectrograms for the three eclipses in the line of contacts, the image of the corona question, a solar radius corresponds to about at the upper and lower edges of the spectrum 50-100 A. The average intensity of photo- is not affected by the motion of the moon spheric continuum over a 100-200 A band is
THERMODYNAMIC STATE OP THE CHROMOSPHERE 37 very nearly equal to the intensity at the central one near X7000, were obtained in this way. wavelength of the band, and we may neglect These curves served to determine coronal the change in wavelength in the displaced brightness, E\(R), as a function of R. The fact images. that E\(Ri)/E\(R2) proved to be the same at Let F\(R) represent the surface brightness of these two wavelengths furnished a check on the polar corona at wavelength X and distance the accuracy of the calibration technique. R from the center of the disk, and let y be the Characteristic curves at other wavelengths were coordinate of a point in the coronal image then derived from the measured E\(R). The measured along the direction of dispersion; shapes of the curves and displacement in in- then the intensity in the coronal image at a tensity with wavelength agreed with those ob- point beyond the limb ±90° to the line of tained from the standard-lamp exposures re- dispersion is peated after eclipse, except for the very toe of the characteristic curves. The double images f° (1) of the spectrum also gave results consistent with the standardization from the coronal Fortunately, eclipse observers customarily image. The absolute intensity scale was fixed take one or more long exposures of the coronal by the standard-lamp exposures. spectrum during midtotality, and the three In 1932, the photospheric spectrum and a eclipses under discussion were no exception. step-wedge sensitometer provided the stand- Within the limits of reciprocity failures, these ards. The sensitometer exposures gave the different exposure times furnish an additional shapes of the characteristic curves, and the check on the standardization. photospheric spectrum was used to obtain Direct standardizations.—The primary stand- absolute and relative intensities. The absolute ardizations of the 1952 and 1936 spectrograms intensities were based on the assumption that were based on exposures from tungsten standard the extreme edges of the photospheric disk radi- lamps. The standardizing exposures gave char- ate as a black-body of temperature 4700°. The acteristic curves on an absolute intensity scale relative intensity corrections were based on the as well as the usual corrections for differential assumption that the continuous spectrum of the apparatus and film functions. Corrections for chromosphere and corona corresponds to a atmospheric extinction were measured at the black-body at 5700°. As mentioned above, eclipse sites. In neither case were the stand- this procedure automatically corrects for all ardizing exposures at the eclipse site completely differential effects in the observing equipment successful and auxiliary standardizations and for atmospheric extinction. proved to be necessary. We turn now to an attempted check of the The auxiliary standards for 1936 received direct standardizations by using the coronal exposures and processing different from those of images as secondary standards. the eclipse films. These tests agreed reasonably Coronal standards versus direct standards.—At well with results from the original standardizing the 1952 and 1936 eclipses the atmosphere was exposures, except for relatively minor differences clear, and we have little reason to expect trouble in detail. if we use the corona as a standard source. Each 1952 spectrum possessed as auxiliary However, in 1932 the eclipse was observed standards the filtered image of a step wedge, through thin clouds, and we cannot hope for taken simultaneously with the spectrogram. completely reliable measures of intensities in A beam splitter formed two images of known the coronal continuum. The original standard- intensity ratio for each spectrum. In addition, ization matched the color of the combined the calibration by means of a standard lamp chromospheric and coronal continuum near the was repeated after the return of the expedition. line of contacts to the photospheric curve. The step-wedge exposures were used to con- This same standardization, however, does not struct characteristic curves at the wave- match the polar corona to the photospheric length of the transmission band of the filter. curve. Similarly, when we use the polar corona Curves of two wavelengths, one near A5000 and as the standard, the equatorial regions do not
SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS 38 match. The discrepancy undoubtedly arises and Dombrovsky (1941), and Zonn (1937), from the cloud cover, and raises serious doubts when averaged together, agree well with the about the justification of using the continuum 1936 curve from the flash spectrograms. The in either the original standardization or the 1936 and 1952 curves agree satisfactorily with proposed standardization against the corona. the maximum and minimum coronal models, This doubt does not necessarily imply that line respectively. However, the 1932 curve diverges intensities are similarly affected since they are widely from any of the accepted models. obtained by integration above the continuous Absolute intensities of E*(R) at X 4700 and background. B=l.l are given in table 1. Two values are In spite of the doubts about using the corona given for 1932. The higher value corresponds as a standard in 1932 it is of interest to carry TABLE 1.—Absolute intensities (Log E\ (R)* for X4700 through the reduction to see its effect on the and R= 1.1) data. Max. Min. 1932 1936 1952 Figure 1 contains plots of E\(R) at\4700 for various representations of the polar corona 12.72 12.18 12.05 12.75 12.60 12.48 normalized to give a common value at R= 1.05. •Intensity units are erg sec-' dX—1 A for the radiation in all directions The curves for sunspot maximum and mini- from a 1 cm slice of atmosphere. mum were derived from van de Hulst's (1953) coronal model. The curves labeled 1932, 1936, to the short exposures of theflashspectrum and and 1952 are those obtained from the flash the lower values to the long exposure during spectrograms. The observations of the corona midtotality. Both the 1936 and 1952 results in 1936, by Bugoslavskaya (1941), Vsessviatsky are independent of exposure time. Again, the 2.0 1.6 4-1932 1.2 •MAX. 0.8 1.2 1.4 t.6 FIGURE 1.—Radial brightness distribution for polar corona on slitless spectrograms.
THERMODYNAMIC STATE OF THE CHROMOSPHERE 39 1936 and 1952 results are reasonably consistent against the corona eliminates most of the differ- with the coronal models, whereas the 1932 ences between the 1932 data and the data from results lead to inconsistencies on different the other eclipses. In the subsequent discus- exposures. sions, unless otherwise stated, all references to Since the 1952 eclipse occurred two years be- 1932 data refer to the standardization against fore sunspot minimum, the absolute intensities the corona. seem to be too high by 0.2 to 0.3 relative to the coronal models. The 1936 values appear to Data be within 0.1 of the expected values. Definition of h=0.—The zero point of the The 1952 spectrograms show essentially the height scale in 1952 was defined as the height same intensity distribution with wavelength in where T, 7O O=1 for a tangential ray. If the the coronal continuum as the continuous spec- solar atmosphere near the limb is assumed to trum of the integrated photospheric disk. Both be isothermal and of constant scale height, the 1932 and 1936 spectrograms show a slight ^4700=1 for a tangential ray when d?E/dh>=0. ultraviolet deficiency. Hence, on a plot of E versus height, A=0 at From the above discussion it is evident that the point of inflection. To a good approxima- the calibrations of the 1952 and 1936 eclipse tion, the same is true on a plot of log E versus spectrograms are reasonably consistent with the height. coronal models, whereas the 1932 calibrations The continuum data at X4700 for the three are inconsistent. It is not clear how much of eclipses are plotted in figure 2. The solid the difficulty in 1932 to attribute to interference curve below 200 km is the curve obtained for by clouds and how much to the original cali- a black-body at X4700 and T 47O O=«~' /80 , where brations. Thus, the 1952 and 1936 results can k is measured in km. The 1932 and 1936 be compared with some confidence, but any curves were adjusted horizontally to place the discrepancies that appear in the 1932 data of point of inflection at h=0. The uncertainty Cilli6 and Menzel (1935) must be regarded with in this assignment oi h=0 does not appear to suspicion. be more than ±50 km. The 1936 and 1932 spectrograms were re- Tabulations of 1936 data.—Tables 2-6 contain standardized against the 1936 and the sunspot the measured intensities of chromospheric lines minimum coronal models. For the 1936 spec- from three separate spectrographs. The in- trograms this restandardization required only tensity units are ergs sec"1 for the radiation slight modifications of the intensities in the in all directions from a slice of atmosphere ultraviolet. For the 1932 spectrograms, how- 1 cm wide, bounded radially by the moon on ever, a complete revision of the standardization one side and extending to co on the other. In was necessary. New microdensitometer trac- general, the blended lines are listed under the ings were made at regions of the limb selected element that comes earliest in the alphabet. as having no signs of abnormal activity. The Thus, most of the blends with Ti lines are 1936 data are tabulated in the following section. listed under preceding elements in the tables. However, in view of the uncertainty in the 1932 We have made exception of lines occurring standardization we have not tabulated the data. in close multiplets, e. g., the blended line of As we have pointed out above, the standard- Fe and Mg at X5167 is listed with the other ization of the 1932 spectrograms against the two lines of the Mg triplet. Lines measured corona is not necessarily better than the original on the spectrograms from the three spectro- standardization. Indeed, in view of the vari- graphs are tabulated separately because the able cloud cover, the reverse is more likely to mean heights of the exposures were not the be true. Nevertheless, the standardization same.
40 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS 4.9 4.0 X 1952 \, \\ 0 1956 . 1992 9.5 5.0 2.5 2.0 1.5 I -—\\\ 1.0 0.5 1 O 2000 4000 HCI0HT (KM.) FIGURE 2.—Continuum intensity near solar limb at X4700. With few exceptions the line intensities can our intensities against Hemmendinger's (1939) be adequately represented by an exponential for the same region on the limb. The plotted emission gradient of the form, points define a straight line of slope 0.85, and the two scales give equal intensities at log E=Eo e-
THERMODYNAMIC STATE OF THE CHROMOSPHERE 41 TABLE 2.—Hydrogen, helium, strontium, and calcium spectrograph No. S Height Wave- Line tfX10*cm~i identity X 290 1000 1720 2440 3140 3850 4650 5250 5950 6660 7370 8080 8800 4340 HT 4,77 4.57 4.16 3.63 3.39 2.77 2.25 1.41 4101 H< 4.86 4.47 4.12 3.72 3.22 3.09 2.70 2.27 L17 3889 H8 4.96 4.66 4.16 3.76 3.30 2.87 2.62 L88 3836 H9 4.69 4.29 3.68 3.16 2.84 2.68 L62 3798 H10 4.54 4.14 3.53 2.94 2.67 L70 3771 Hll 4.51 4.02 3.35 2.72 L82 3750 H12 4.47 3.96 3.27 2.71 1.82 3734 H13 4.43 3.92 3.25 2.59 1.97 3722 H14 4.36 3.76 3.17 L88 3712 H15 4.13 3.64 2.97 L83 3704 H16 4.06 3.45 2.83 LOS 3697 H17 a 97 3.36 2.0 3692 H18 3.88 3.28 2.0 3687 H19 8.87 3.18 2.2 3683 H20 a 76 3.16 2.0 3679 H21 3.88 3.15 2.4 3676 H22 3.80 3.06 2.4 3674 H23 3.65 2.96 X2 3671 H24 3.62 2.87 2.4 3660 H25 3.63 2.97 XI QAJUZ OuuO H26 3.38 2.78 XO 3666 H27 3.35 2.78 LO 4713 He I 2.61 2.48 2.14 0.8 4472 He I 4.18 4.00 3.72 3.43 3.09 2.76 2.67 LOO 0.99 4026 He I 3.47 3.05 2.86 2.61 2.46 2.24 0.90 4388 He I 2.12 4686 Hen 2.10 1.68 0.7 4078 Srn 4.62 4.05 3.35 2.77 2.53 2.31 1.57 4215 Srn 4.39 3.89 3.05 2.43 2.14 2.09 L75 3969 Can 4.85 4.65 4.46 3.98 3.68 3.47 3.02 2.73 2.54 Z28 a 99 3034 Can 4.91 4.76 4.62 4.13 3.90 3.75 3.30 2.91 2.75 2.40 0.07 4227 Gai 4.10 3.43 2.54 X44 3701 Ca i-Tl n 3.97 3.18 2.6 improvement over the value 0.85 obtained 1932 data.—The restandardization of the 1932 from comparison of our data with Hemmen- spectrograms, as we indicated in the previous dinger's. section, leads to results considerably different The difference in reduction methods probably from the data published by Cilli6 and Menzel accounts for the fact that our data show better (1935). Table 6 contains the new fi's for a agreement with Menzel's results than with limited number of lines. They are systemati- Hemmendinger's. In our reduction and in that cally greater than Cilli6 and Menzel's values for of Menzel, the density profiles were replotted the same lines by about a factor 1.5. as intensity profiles on an enlarged scale before the areas under the profiles were measured. Discussion Hemmendinger used a mechanical device to The 1952 data are more complete with regard measure the intensities directly from an inte- to wavelength coverage and height resolution gration of the density profiles. The integrating than are those of 1936 or 1932. However, the device he used introduced a possible additional 1936 and 1932 data are of great value for indi- source of systematic errors in the results. The cating possible variations during the sunspot differences that are present in our characteristic cycle and in indicating the uncertainties in the curves appear to be both too small and in the 1952 data. In the introduction to this paper wrong direction to explain the differences in we posed two problems: Do observations of the the data. It seems unlikely that the inte- chromospheric spectrum over a narrow sector grating device used by Hemmendinger would of the limb at a given eclipse suffice to deter- lead to errors as large as those indicated. Thus, mine the properties of the average chromo- it appears that a combination of the reduction sphere at the time of the eclipse? If so, does techniques and photometric standardizations this average change with time? Abundant rather than a single phase of either operation evidence suggests an affirmative answer to the accounts for the differences in the two sets of first question. Hemmendinger measured line data. intensities and emission gradients at 14 regions
42 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS TABLE 3.—Metals spectrograph No. S on the solar limb at the 1936 eclipse. Nine of Wave- Height these regions were essentially the same in both length X Line identity 0X1O1 line intensities and emission gradients. All of 290 1000 1720 2440 the remaining regions showed visible promi- 3962 All 3.77 3.08 2.2 nences extending above the chromosphere. At 3044 4554 All Ban 3.76 4.22 3.42 2.22 3.2 the 1952 eclipse, Athay, Billings, et al. (1954), 4565 CTI 2.62 measured line intensities and emission gradients 4344 Cr l-Ti ii 3.49 4290 Cr l-Ti n-Ca I 4.19 2.99 3.9 at two regions on the Limb and found no discern- 4275 Cr i-Ti i 2.61 3.1 4254 Cn 3.82 3.78 2.88 2.9 ible differences. Houtgast (1953) determined 4588 Cm 3.14 45S1 Fe i-Ca i 2.97 emission gradients at the 1952 eclipse by com- 4556 4534 " -Tin 3.70 4.19 3.19 2.59 3.1 bining data from several points on the limb to 4415 4408 " -Sen || -Tin 3.79 3.29 2.24 5.0 obtain the necessary dispersion in height. His 4405 4384 " 3.66 3.93 2.55 2.98 3.6 3.0 results agree quite well with those of Athay, 4375 4326 " -Scn-Yn " -Nil 4.10 3.91 2.74 2.79 4.4 3.6 Billings, et al. (1954). 4294 4272 '.'. ' T i n 3.83 3.66 2.78 2.35 3.4 4.3 At most eclipses there are outstanding regions 4360 4250 3.20 3.52 2.00 2.09 3.2 4.6 of peculiar emission characteristics. Such re- 4072 4064 •• 3.60 3.66 2.45 2.57 3.7 3.6 gions have been variously referred to as "hot 4045 3914 • -Tin 3.75 3.86 2.81 3.03 3.0 2.7 spots," "excited regions," and "active regions." 3886 3879 • -Lan | -Vn 3.67 3.65 2.74 3.0 However, in all reported cases of this type the 3860 3856 ' -Sin 3.83 3.48 3.16 2.77 2.2 2.3 anomalous characteristics are concentrated in 3840 3826 • • -CN 3.40 2.67 2.82 1.9 regions l°-2° wide located over sunspot and 3824 3820 Fei Fe I-He I 3.41 3.57 2.66 3.05 2.4 1.7 plage areas. At any one eclipse, they make up 3816 3795 Fei •* 3.24 2.66 only a very small percentage of the chromo- 3767 3764 3.47 3.35 2.51 2.78 3.1 1.9 sphere. Thus, we conclude that measurements 3746 3720 3.78 3.11 3.21 2.2 of the chromospheric spectrum at one region of 3737 3728 " -Can-Nli " -Zrn 3.94 3.74 3.45 1.6 the limb are capable of giving a fair picture of 4629 4584 Fe n-Ti I 3.81 4.03 2.69 4.3 the chromospheric spectrum. It should be 4576 4559 2.88 2.13 4.1 pointed out that physical limitations on the " -Cm 3.38 4550 4523 " -Tin " -Tii 4.28 3.69 3.25 2.32 3.2 size of microdensitometer slits and on the re- 4520 4515
THERMODYNAMIC STATE OF THE CHROMOSPHERE 43 TABLE 4.—Spectrograph No. 6 Height Wave- Line identity 0 X 10* length X 560 1360 2140 2900 3700 4480 5220 6060 6800 7600 6563 Ha 5.34 5.14 4.76 4.24 3.96 1.20 4861 H/J 5.52 5.24 4.93 4.75 4.41 3.97 3.35 2.96 2.66 2.07 1.37 6876 He i 5.19 4.86 4.44 4.19 3.91 3.60 3.13 2.78 2.43 2.17 0.99 5016 2.71 2.69 2.26 1.91 0.88 4922 2.68 2.46 0.6 4713 2.40 2.37 2.14 2.16 0.4 4686 Hen 1.85 1.76 1.62 0.4 4934 Ban 3.55 2.96 1.7 5328 Cr i-Fe i 3.40 1.89 4.3 5208 Cri 3.37 2.02 3.9 5206 " 3.58 2.04 4.3 5204 " 3.12 1.72 4.0 5169 Fe i-Fe II 3.92 2.89 2.42 1.97 2.1 4921 Fel 2.70 1.46 3.5 4919 " 2.26 5317 Fen 3.54 1.97 4.5 5018 Fe n 3.97 2.78 1.93 2.9 4924 " 3.78 2.55 3.5 5184 Mgi 4.24 3.65 2.92 2.42 2.00 1.75 5173 «< 3.97 3.44 2.55 2.30 1.66 1.89 5167 " -Fei 3.87 2.95 2.03 1.01 3.04 5896 Nai 4.01 3.13 2.22 1.59 2.34 5889 Nai 4.08 3.33 2.43 ZOO 2.25 14.0 _ Ul 13.0 _ o o 12.0 _ 12.0 13.0 14.0 LOG E (AMO) FIGURE 3.—Correlation diagram for our measures of line intensities versus Hemmendinger's.
44 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS 14.0 13.0 12.0 12.0 13.0 14.0 LOG E (AMO) FIGURE 4.—Correlation diagram for our measures of line intensities versus Menzel's. between polar and equatorial regions, give TABLE: 6.—Emission gradients, 19SS marked indication of variations. In the chro- Wave- Line 0X1O» mosphere we do not have sufficient data in the length X identity polar regions to compare the polar and equa- 3722 H14 1.8 TABLB 5.—Spectrograph No. 6 3712 H15 1.6 3704 H10 1.9 3097 H17 2.0 Wave- Height 3692 H18 1.9 length Line 0X1O» 3687 H19 1.9 X identity 3683 H20 2.0 380 1060 1780 2520 3220 3679 H21 2.1 3676 H22 2.2 3674 H23 2.3 7005 He I 4.29 4.06 3.82 3.51 3.27 0.85 3671 H24 2.3 0078 He I 3.75 3.47 3.09 LOO 3669 H25 17 3068 H26 2.2 3600 H27 2.2 3040 Ho» 2.3 torial regions. Hence, though we may use the 4713 4020 He I He i 1.3 1.4 4686 0.9 emission gradients as possible indicators of 4072 Hen Fei 2.6 variability, we must also compare absolute and 4004 4045 3.0 2.8 relative intensities. 3764 3746 3.1 2.2 3720 2.1 Emission gradients.—We obtain the most sat- 4078 Srn 1.7 3759 2.0 isfactory comparison of emission gradients by 3761 Tin " 2.0 3714 3.8 comparing the 1936 and 1932 gradients with 3710 Vn Yn 4.3 those of 1952, since they have more lines in
THERMODYNAMIC STATE OF THE CHROMOSPHERE 45 common. The emission gradients for the metal systematically high in 1936 and low in 1932. lines at the 1952 eclipse are not yet available in Then* results represented accurately the avail- the literature, but will soon be published by able data at that time. The differences between J. B. Zirker. Figures 5 and 6 exhibit the indi- our measures of the emission gradients in 1936 cated comparisons. In both cases straight lines and Hemmendinger's, which were used by passing through the origin with slope 1.0 are Athay and Thomas, arise from the greater adequate representations of the plots. The height range in our data. A given spectrogram faint metal lines have the largest /3's. Because usually shows systematic errors in intensities of the faintness of these lines, the probable resulting from such effects as instrumental vi- errors in the /3's are relatively large, as evi- brations, seeing, focus, exposure time, etc. denced by the increased scatter for large 0'a Hemmendinger's reduction was restricted to in figure 5. optical densities below about 1.3. By using a Figures 5 and 6 show no indication of change more sensitive microdensitometer we extended in the emission gradients through the sunspot the reductions to densities of about 2.5 and cycle. This is contrary to the results reported were thus able to measure lines at lower heights. by Athay and Thomas (1956), who found that The increased height range in our data gives a the emission gradients, relative to 1952, were corresponding increase in the accuracy of the 4 - ^ 3 CM •n — CO o 2 _ I _ 1 2 3 4 8 fiXIO (1936) FIGURE 5.—Correlation diagram for 1936 0*8 versus 1952 0's.
46 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS 3 X p X I0 8 (1932) FIGURE 6.—Correlation diagram for 1932 /5's versus 1952 /3's. emission gradients. The 1932 emission gradi- Balmer decrement. The observed decrements ents used by Athay and Thomas were those at the three eclipses are shown in figure 7 at as given by Cillie" and Menzel, which are, as we nearly a common height as the data allow. have noted, systematically low. For comparison, we have also plotted the decre- Relative intensities.—The relative intensities ment given by Cillie' and Menzel. Although of the high-order Balmer lines provide a useful their decrement is markedly flatter than either indicator of changes in chromospheric structure. of those for 1936 and 1952, the restandardiza- The observed intensities are controlled to a large tion of the 1932 spectrograms gives a decrement extent by the optical thickness of the chromo- in fairly good agreement with those of 1936 sphere resulting from absorption by the second and 1952. The lines Hll to H14 appear to be quantum level of hydrogen. In the chromo- stronger, relative to the higher order lines in spheric regions where the hydrogen emission is 1932, than in either 1952 or 1936. However, concentrated, the kinetic temperature is near all of these lines are blended with metal lines 6000° (Athay, Menzel, Pecker, and Thomas, with steeper emission gradients, and the appar- 1955). At this temperature the population of ent strengthening of the hydrogen lines probably the second quantum level varies rapidly with arises from the metal lines because of the some- temperature and density, and we may expect what lower height. As far as relative intensities changes in the model to be reflected in the are concerned, the Balmer decrements are con-
THERMODYNAMIC STATE OF THE CHROMOSPHERE 47 sidered to be in satisfactory agreement at the extend to greater heights. There is no evidence three eclipses. We shall postpone discussion of that line intensity changes significantly with the absolute intensities until the following sec- excitation potential from one eclipse to another. tion of this paper. Absolute intensities.—The data in figures 7, A still more sensitive indicator of changes in 8, and 9 indicate that the absolute intensities chromospheric structure is given by the relative in 1952 are systematically higher in 1936 and intensities of helium, hydrogen, and metal lines. 1932, by as much as 0.5 to 0.6 in the logarithm. Since the line at X4686 from He n differs widely The continuum data in figure 2, however, in excitation energy from some of the metal show the 1952 continuum intensities near h=0 lines, even slight changes in excitation tem- to be about 0.25 above the 1936 intensities, and peratures would produce marked changes in about 0.5 above the 1932 intensities. The con- relative line intensities. Figure 8 is a plot of tinuum intensities above 1,000 km show still such lines selected from the data for the three different effects. At these heights, however, eclipses. Figure 9 exhibits some of the stronger the corona and stray light in the spectrographs lines from the 1936 and 1952 eclipses, which contribute strongly to the observed emission, 15.4 V 15.0 , 1952 (960 KM) 1936 (1000 KM) 14.6 ,1932 (790KM) 14.2 J932 (CSM, 670 KM) c 14 o o 13.8 13.4 13.0 12 6 12 16 20 24 28 32 n FIGURE 7.—Balmer decrements for 1952, 1936 and 1932 eclipses.
CONTRIBUTIONS TO ASTROPHYSICS VOL.2 15 14 13 12 IOOO 2000 3000 IOOO 2000 3000 14 . 14 4072 Ft 2 4686 Hen 13 IOOO 2000 3000 0 IOOO ,2000 IOOO 2000 3000 HEIGHT (KM.) . 1932 OI936 XI9S2 FIGURE 8.—Plots of log E versus height for lines of intermediate strength. and we cannot expect agreement at the three factor two between the line and continuum in- eclipses. If we force the continuum intensities tensities in 1936. This discrepancy, of course, to agree at k=0, the line intensities for the may be simply a photometric difficulty. If it is 1952 and 1932 eclipses are also in good agree- real, it represents a change of about the same ment, but the line intensities for the 1936 magnitude, but in the opposite sense to the eclipse are relatively weak by about 0.3. The changes in the corona. Much more reliable 1952 and 1932 eclipses both occurred near sun- absolute intensities are needed before definite spot minimum, and it is reasonable to suppose conclusions can be drawn. However, it seems that the chromosphere would be relatively un- clear from the above data that chromospheric changed. The fact that both lines and con- changes in brightness are of no greater magni- tinuum can be brought into agreement for the tude than coronal changes, and there is no evi- 1932 and 1952 eclipses suggests strongly that dence for significant changes either in excitation the observed differences in absolute intensity conditions or in emission gradients. result from photometric difficulties rather than If we grant that the absolute intensity scales from real differences in the chromospheric in 1932 and 1952 should be adjusted by a rela- emission. tive amount of 0.5, it is still not clear just how From the above discussions it seems evident this adjustment should be made. The data in that the only possible indication of significant table 1 indicate that the 1952 intensities are too changes in chromospheric emission over the high by 0.2 to 0.3. In the photograph of the sunspot cycle is the apparent discrepancy of a 1932 corona published by Moore (1932), the
THERMODYNAMIC STATE OF THE CHROMOSPHERE 49 16 16 H Call 15 15 14 14 13 13 12 I 12 I 2000 4000 £000 8000 2000 4000 6000 8000 15 H c l 15 5184 Mg 14 14 13 13 12 12 I I 2000 4000 6000 8000 0 2000 4000 6000 8000 HEIGHT (KM.) O |936 X 1952 FIGURE 9.—Plots of log E versus height for strong lines. polar corona at the pole that we used in stand- admit strong variations in both emission gra- ardizing the spectrograms is much brighter dients and relative intensities. The absence of than the corona over the opposite pole. Thus, such effects between the 1936 and 1952 data the most reasonable adjustment of intensity suggests that the 1932 data obtained from the scales seems to be an average of the 1932 and restandardization are the more reliable. 1952 absolute intensities. On this basis, no The authors are indebted to J. B. Zirker for adjustment is necessary for the 1936 intensities providing the metal line data from the 1952 if we use the photospheric continuum near the eclipse prior to publication, and to Dr. R. N. limb as reference. Thomas for stimulating interest in the problem. The possibility remains that the original 1932 This work was supported in part by the data of Cilli6 and Menzel are more reliable than Office of Naval Research, carried out in co- the data we have used, in which case we must operation with the Naval Research Laboratory,
50 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS and in part by the Air Force Cambridge Re- ClLLIE, C . G., AND MENZEL, D . H . search Center, Geophysics Research Direc- 1935. Harvard Obs. Circ, No. 410. DUNN, R. B. torate, through Contract AF 19 (604)-146 with 1956. Astron. Journ., vol. 61, p. 3. Harvard University. HEMHENDINOER, H. 1939. Dissertation, Princeton Univ. References HOUTGAST, J. ATHAT, R. G. 1953. Convegno Volta 1952 Roma, Accademia 1953. High Altitude Obs. Techn. Rep., July 13, Nazionale Lincei, p. 68. 1953. HULST, H . C. VAN DE ATHAT, R. G.; BILLINGS, D. E.; EVANS, J. W.; AND 1953. In Kuiper, ed., The sun, p. 207. ROBERTS, W. O. MENZEL, D. H. 1954. Astrophys. Journ., vol. 120, p. 94. ATHAT, R. G.; MENZBL, D. H.; PECKEB, J.-C, AND 1931. Publ. Lick Obs., vol. 17, p. 1. THOMAS, R. N. MOORE, J. H. 1955. Astrophys. Journ., Suppl. No. 1, p. 505. 1932. Publ. Astron. Soc. Pacific, vol. 44, p. 341. ATHAT, R. G., AND THOMAS, R. N. VSESSVIATSKT, S. K., AND DOMBBOVSKT, V. A. 1956. Astrophys. Journ., vol. 123, p. 309. 1941. In Report of Soviet Expedition 1936, vol. 2, BUOOSLAVSKATA, E . J . p. 104. 1941. In Report of Soviet Expedition 1936, vol. 2, ZONN, W. p. 74. 1937. Acta Astron., ser. A, vol. 3, p. 135. Abstract Chromospheric line and continuum intensities obtained from jumping-film observations of the 1952, 1936, and 1932 eclipses are compared for the purpose of indicating changes in chromospheric structure during the sunspot cycle. The 1936 and 1932 spectrograms are restandardized against the corona as a standard source, and the zero-points of the height scales are redefined to be consistent with the 1952 height scale. For the 1936 spectrograms, the restandard- ization requires only slight modification of the original standards, but for the 1932 spectrograms a complete revision is required. The tabulated data include chromospheric line intensities and emission gradients for the 1936 eclipse and emission gradients of a few selected lines for the 1932 eclipse. No evidence is found for significant changes in emis- sion gradients and relative line intensities during the sunspot cycle. Absolute intensities of chromospheric lines in 1936 appear to be weaker, relative to 1952 and 1932, by a factor two. This apparent change in emission may result from photometric uncertainties; however, more accurate absolute intensity measurements are necessary before definite conclusions can be stated.
You can also read