1 INTRODUCTION

The aurora is one of the important phenomena in the solar-terrestrial environment. It is regarded as an indicator in the solar wind-magnetosphere coupling process in polar ionosphere. Auroral emissions are excited by collisions between precipitating electrons and neutral particles in atmosphere. The photon is generated by collisional deexcitation after bound electrons transit from ground state to excited state (Jones, 1974). There are three major spectral lines in the visible band (427.8, 557.7 and 630.0 nm), which play an important role in auroral emissions (Chamberlain, 1995). The generation of green aurora (G, 557.7nm) is associated with energy level transitions from 1S to 1D in atomic oxygen O and the life-time in the second excited state 1S is 0.8 s. Red aurora (R, 630.0 nm) is generated by energy level transitions from 1D to ground state in atomic oxygen with a life-time of 110 s. Blue-violet aurora (V, 427.8 nm) is excited by the first-negative band system of nitrogen molecules N2 (N+). Owing to the different generations of 427.8, 557.7 and 630.0 nm auroras, the height of the maximum luminosity is not the same. The height of red aurora is highest (~200 km), since the excited O (1D) would be subject to collisional deactivation at lower altitude due to the long life-time of O (1D) (Meng et al., 1991). The density of atomic oxygen decreases quickly and that of nitrogen molecules increases at lower altitude. As a result, the height of maximum luminosity (~150 km) in green aurora (557.7 nm) is higher than that (~120 km) in blue-violet aurora (427.8 nm).

Many scholars have made useful researches about correlation between auroral emission intensities at dif-ferent wavelengths by analyzing the spectra of precipitating particles. An earlier study by Eather and Mende showed the relevant significance between latitudinal distributions of auroral intensities (427.8, 557.7 and 630.0 nm) and particle fluxes (Eather, 1969; Eather and Mende, 1972). Later, Gattinger and Jones (1972) found that the I_630.0/I_427.8 and I_557.7/I_427.8 ratios decrease when the I_427.8 emission intensity increases. Jones et al. (1987) found that the average energies inferred from the radar electron density profiles and the I_630.0/I_427.8 ratio are in good agreement, but there is little variation of the I_557.7/I_427.8 ratio. Recently, Dashkevich et al. (2006) obtained the dependences of the I_630.0/I_427.8 and I_557.7/I_427.8 ratio on the I_427.8 emission intensity using more than 3800 measurements. It is found that the I_630.0/I_427.8 ratio decreases from 2 to 0.4 when the I_427.8 emission intensity increases from 0.1 to 3 kR. In the I_427.8 emission range 0.1~1.8 kR, the I_557.7/I_427.8 ratio tends to increase and takes the values 4.2~6.4. These results represent a relationship between auroral electron fluxes and spectroscopic emission features in aurora (Rees and Luckey, 1974). Rees and Roble (1986) and Christensen et al. (1987) had given approximate equations for auroral emission intensity as a function of electron spectrum and the I_630.0/I_427.8 ratio as a function of the characteristic energy.

YRS is located at the cusp latitude, which is suitable to observe dayside aurora. Hu et al. (2009) examined the synoptic distribution of dayside aurora and divided into 4 categories using the observations from three-wavelength (427.8, 557.7 and 630.0 nm) all-sky imagers (ASIs) at YRS. Dayside auroras and dynamic processes in magnetospheric boundary layer are closely related. Therefore, the source region should be determined firstly in the study of dayside aurora. Newell et al. identified the source region of auroras at low altitude using precipitating electron spectra observed from DMSP satellites (Newell and Meng, 1988; Newell et al., 1991a; Newell et al., 1991b). There are three major source regions of dayside aurora at low altitude: (1) LLBL, located between 74°and 77°MLAT which is higher near magnetic noon and lower at dawn and dusk, total energy flux of precipitating electrons is larger than 6 ×10¹⁰ eV·(cm²·s·sr)^-1, the energy fluxes of 2 keV and 5 keV electrons are less than 10⁷ eV·(cm²·s·sr)^-1, average electron energy is between 220 and 600 eV, number flux of electrons is about 10⁹ (cm²·s·sr)^-1 (Newell and Meng, 1988); (2) Mantle, located in poleward of cusp or dayside auroral oval (~77°MLAT), a general tendency for the density and average ion energy is to decline with increasing latitude, the average energies of electron and ion are low (about a few times of 100 eV) (Newell et al., 1991a); (3) BPS, electron average energy has a magnitude of keV, the energy fluxes of 2 keV and 5 keV electrons are larger than 10⁷ eV·(cm²·s·sr)^-1 (Newell et al., 1991b).

Because DMSP is a polar-orbiting satellite, there are only a few simultaneous observations by DMSP satellites and YRS at noon and in the postnoon sector. It is hard to obtain the precipitating particle spectra for any dayside aurora. Thus, it is necessary to build a parameter model between emission intensities in dayside aurora and precipitating electron spectra. Xing et al. (2013) had built a model of auroral emissions and particle precipitation near magnetic noon at YRS using ASIs (427.8 and 630.0 nm) and DMSP particle spectra data. Auroral arc is one of typical forms of discrete aurora distinguished by Hu et al. (2009) in the prenoon and postnoon sectors (Yang et al., 2000; Sandholt et al., 2002). A stable auroral arc is elongated in longitude ( > 100 km) and narrowed in latitude (~18 km). The intensity of the green line (557.7 nm) is roughly proportional to that of the emission at 427.8 nm and is a good indicator of electron of energy flux (Paschmann et al., 2002). Dayside auroral arcs at 557.7 and 630.0 nm are bright, but the I_427.8 emission intensity is so weak that we use I_557.7 instead. Consequently, we would use the I_557.7 and I_630.0 emission intensities to analyze the spectra of precipitating electron in this study.

In this paper, the dependence of I_557.7/I_630.0 ratio on the I_557.7 emission intensity was studied using dayside auroral arc images (0600-1800 MLT) from multiple-wavelength ASIs at YRS. And then, we built a model of auroral arc emissions and particle precipitation by analyzing the corresponding source region by means of contrasting simultaneous observation from DMSP. Using this model, we determined the spectra and source region of dayside auroral arc without DMSP data and discussed the MLT distribution of source regions.

2 DATA

The auroral data used in this study were obtained from ASIs with 557.7 and 630.0 nm filters from December 2003 to February 2007 at Chinese Yellow River Station (YRS), Ny-Ålesund, Svalbard. YRS is located at 78.92°N, 11.93°E, and corrected geomagnetic latitude 76.24°, where MLT≈UT+3 h (Qiu et al., 2013). The optical system at YRS contains three identical all-sky cameras (ASC), which were developed to measure the photoemissions at 427.8, 557.7 and 630.0 nm (Hu et al., 2005). The CCD (Charge-Coupled Device) camera uses a 512×512 square pixel array, and the center of the field of view is geographic zenith. ASIs are taken with a 7 s exposure time at 10 s cadence. ASIs are taken according to zenith and azimuth angle using a fish-eye lens, so that any pixel in an ASI has the same solid angle. Therefore, auroral emission intensities were corrected with zenith angle based on the model of Yang et al. (1997). We selected these auroral arcs which were bright both at 557.7 and 630.0 nm ASIs in this study. In total, there were 22882 auroral arcs observed in 44 days.

For each arc ASI, the emission intensity variation curve can be obtained as a function of zenith angle by extracting along magnetic meridian (Fig. 1). Fig. 1a shows an arc ASI taken at 1359 UT (1659 MLT) on December 24, 2003. The white line is along magnetic meridian and the field of view of ASC is in the range of the circle with diameter of this white line. From Fig. 1b, we can see a peak marked by a vertical dashed line, which represents an auroral arc. The value (4.7 kR) of the peak is defined to be the emission intensity of this arc in Fig. 1b. The emission intensity of any other arc was measured by the same method either in 557.7 or 630.0 nm ASI.

The Defense Meteorological Satellite Program (DMSP) is a collective name of a series of polar-orbiting meteorological satellites launched by US Department of Defense. The satellites were launched every two years and named as F (n) one by one. We used the particle spectra data detected from F13-F16. Total number fluxes (J ) and energy fluxes (J E) between 32 eV and 30 keV were measured by SSJ/4 sensors on the DMSP satellites (Hardy et al., 2008). The orbit of DMSP satellite is at a nominal altitude of 830 km with a period of 101 min (Aruliah et al., 2005). We found 17 cases of DMSP passing through 40 auroral arcs between 1320 MLT and 1755 MLT from all of ASIs.

3 RESULT 3.1 Dependence of I_557.7/I_630.0 Ratio on the I_557.7 Emission Intensity

The emission intensity distribution of 22882 dayside auroral arcs and the dependence of I_557.7/I_630.0 on I_557.7 were studied firstly, which could give us an impression of dayside auroral arc. Fig. 2 shows the average emission intensities I_557.7 (a) and I_630.0 (b) of dayside auroral arcs as a function of MLT. The value is marked by "★" at any MLT and error bar is plotted by vertical solid line at the corresponding time. The I_557.7 emission intensity is in the range of 0.1 to 17.3 kR and takes an average value of 2 kR. From Fig. 2a, it is found that the I_557.7 emission intensity has two maximum values. The stronger one (2.9 kR) is near 1500 MLT in hot spot region in the postnoon sector and the weak one (2.2 kR) appears at 0900 MLT in warm spot region in the prenoon sector. From Fig. 2b, the range of I_630.0 emission is from 0.2 to 3.9 kR, much smaller than that of I_557.7. The average emission intensity I_630.0 is weaker than I_557.7 having a value of 0.9 kR. Comparing with the distribution of I_557.7, the intensity peak of red auroral arcs (630.0 nm) is located near magnetic noon and the average emission intensity I_630.0 is weaker in the prenoon and postnoon sectors. Dayside auroral arcs can be seen clearly both in 557.7 and 630.0 nm ASIs, and the arcs at 557.7 nm are much brighter.

Figure 3 shows a scatter plot of the I_557.7/I_630.0 ratio as a function of I_557.7 emission intensity at different MLT in the prenoon sector. We took logarithm for a comparative analysis. In Fig. 3, the best fit linear regression lines are plotted with solid black lines (equation and correlation coefficient R are labeled on the upper-left corner) and ±σ deviations from linear regression line are plotted with dashed lines. We can see a linear correlation between I_557.7/I_630.0 and I_557.7 in the double logarithmic coordinate system having the correlation coefficient of ~0.9 and the standard deviation of ~0.1. The I_557.7/I_630.0 ratio increases from 0.2 to 9 with the increasing of I_557.7 from 0.1 to 10 kR. The minimum value of I_557.7/I_630.0 is 0.17 in the prenoon sector, the maximum and average value is 23.9 and 2.6, respectively. From Fig. 3, auroral arc of I_557.7/I_630.0 larger than 1 appeared more frequently near magnetic noon, which is consistent with the result from Fig. 1 that maximum intensity of red auroral arcs (630.0 nm) is located near magnetic noon. It is also found that the I_557.7/I_630.0 ratio is smaller for the same I_557.7 emission intensity as it is close to magnetic noon. Fig. 4 shows a scatter plot of the I_557.7/I_630.0 ratio as a function of I_557.7 emission intensity at different MLT in the postnoon sector. I_557.7/I_630.0 and I_557.7 are taken logarithm and the best fit linear regression lines are plotted with solid black lines just like Fig. 3. The result is similar with that in Fig. 3. There is a linear correlation between I_557.7/I_630.0 and I_557.7 with the correlation coefficient of ~0.9 and the standard deviation of ~0.1. I_557.7/I_630.0 increases from 0.2 to 9 with the increasing of I_557.7 from 0.1 to 10 kR. The minimum value of I_557.7/I_630.0 is 0.24 in the postnoon sector, the maximum and average value is 17.1 and 2.0, respectively. The I_557.7/I_630.0 ratio is also smaller for the same I_557.7 emission intensity as it is close to magnetic noon. Consequently, the I_557.7/I_630.0 ratio is relatively small near magnetic noon and relatively large at dawn and dusk.

3.2 Correlation Between Auroral Arc Emission Intensities and Electron Precipitation

The model built by Xing et al. (2013) shows that the I_427.8 intensity is dependent on the total energy flux of precipitating electrons and the I_630.0/I_427.8 ratio is related to the average energy near magnetic noon. Because the intensity of auroral arc at 427.8 nm is weak and the I_557.7 intensity is roughly proportional to I_427.8, our model used I_557.7 emission intensity instead of I_427.8 emission. The generation of aurora is quenched by collisional deexcitation. Aurora emission intensity is large through multiple collisions when the average energy of precipitating electrons is large (Paschmann et al., 2002). Collision would occur frequently as the number flux of precipitating electrons is large that can result in increase of emission intensities. So, total energy flux (the product of number flux and average energy) of precipitating electrons is larger as auroral arc emission intensity increases. In combination with the fact that the typical energies of the primary auroral electrons range between 100 and 150 km, the green line at 557.7 nm dominates the visual appearance of auroral arcs. For very soft electron spectra (~0.5 keV), the red emission (630.0 nm) above 200 km becomes prominent (Rees, 1963). Overall, average electron energy can be represented by the I_557.7/I_630.0 ratio. The average energy of precipitating electrons is larger as I_557.7 emission intensity is larger, otherwise I_630.0 emission intensity is larger.

Figure 5 shows a scatter plot of average electron energy observed from DMSP versus I_557.7/I_630.0 ratio in the double logarithmic coordinate system. The best fit linear regression line is plotted with a solid line. The slope of linear regression line is 0.99 (≈1). From Fig. 5, we found a notable positive correlation (R = 0.82) between average electron energy and the I_557.7/I_630.0 ratio, because higher energy electrons can be precipitated to lower altitude exciting the green line (557.7 nm). The result shows that the fit of the energy-intensity ratio (I_557.7/I_630.0) relationship is good and the average electron energy can be obtained from the I_557.7/I_630.0 ratio. The relation can be given as

(1)

The average energy of precipitating electron E is given by

(2)

where, the unit of average energy E is eV.

From Fig. 2, it is found that the range of green line (557.7 nm) is much wider than that of red line (630.0 nm). The I_557.7 emission intensity is more suitable to represent the total energy of precipitating electron. Fig. 6 shows a scatter plot of total electron energy flux J E versus the I_557.7 emission intensity in the double logarithmic coordinate system. The best fit linear regression line is plotted with a solid line. From Fig. 6, we found a notable positive correlation (R = 0.82) between the log of total electron energy flux log (J E) and the log of emission intensity log (I_557.7). The relation can be given as

(3)

The total energy of precipitating electron J E is given by

(4)

where the unit of total electron energy flux J E is eV·(cm²·s·sr)^-1.

4 DISCUSSION 4.1 Source Region of Dayside Auroral Arcs

We had analyzed the cases of DMSP passing through auroral arcs above YRS in order to determine source regions in the magnetosphere. They have their own characteristics in different source regions. Fig. 7 shows a case of DMSP passing above YRS between 1420-1422 UT on December 29, 2003. ASIs (557.7 and 630.0 nm) and the orbit above YRS are plotted on the left. DMSP satellite F13 came into and out of view above YRS at 1420 UT (1720 MLT) and 1422 UT (1722 MLT), respectively. The orbit of DMSP above YRS is plotted in solid red passing through three auroral arcs marked by "*". The I_557.7 (green) and I_630.0 (red) emission intensity of auroral arcs, total energy flux, average energy and the spectrograms of electron and ion are plotted from top to bottom on the right. Three vertical dashed lines are labeled at the time of DMSP passing through three auroral arcs. Due to the error during coordinate mapping, there would be a little difference (a few to tens of seconds) between the calculation time and the actual value. From Fig. 7, we can see three peaks in I_557.7 emission intensity variation curve corresponding to three arcs. We can also see an inverted "V" in electron spectrogram for each arc. It is found that the total electron energy fluxes of the first two auroral arcs are in the range of 10¹¹ ~10¹² eV·(cm²·s·sr)^-1, the energy fluxes of 2 keV and 5 keV electrons are larger than 10⁷ eV·(cm²·s·sr)^-1, and the average electron energy is 1~2 keV. Through these characteristics, we determined the source region of the first two arcs to be BPS (Newell et al., 1991b). The third auroral arc is located at 78° MLAT in the poleward of ASI. The density and average ion energy decrease with increasing latitude and the average electron energy is low (695 eV). We determined the source region of the third arc to be Mantle (Newell et al., 1991a). In different definitions, the third arc is classified as drapery-like dayside corona by Hu et al. (2009).

The source region of dayside auroral arcs can be not only BPS and Mantle but also LLBL. Fig. 8 shows a case of DMSP passing above YRS between 1149-1150 UT on December 25, 2003. From Fig. 8, DMSP satellite F15 passed through one arc located at 75.6°MLAT in the equatorward of ASI. We can also see an inverted "V" in electron spectrogram for this arc. The energy fluxes of 2 keV and 5 keV electrons are less than 10⁷ eV·(cm²·s·sr)^-1, and the average electron energy is low ( < 600 eV). The source region of this arc is determined to be LLBL based on the characteristics above.

We found 17 cases of DMSP passing through 40 auroral arcs above YRS. Using the same method, we found the precipitating electrons of 33 dayside auroral arcs come from BPS: the I_557.7 emission intensity is in the range of 0.4~5.4 kR; I_557.7/I_630.0 ratio is larger than 1; average electron energy is between 537 eV and 2.4 keV with an average value of 1.1 keV; total energy flux of precipitating electron is larger than 10¹¹ eV·(cm²·s·sr)^-1. It is found that the source region of 6 dayside auroral arcs is Mantle: the I_557.7 emission intensity is in the range of 276~457 Rayleigh; I_557.7/I_630.0 ratio is between 0.4 and 1.2; average electron energy is 214~695 eV with an average value of 467 eV; total energy flux of precipitating electron is less than 1.2×10¹¹ eV·(cm²·s·sr)^-1. We had found only one auroral arc among these 40 arcs produced by precipitating electron from LLBL: the I_557.7 emission intensity is 457 Rayleigh; I_557.7/I_630.0 ratio is 0.6; average electron is 417 eV; total energy flux is 1.2×10¹¹ eV·(cm²·s·sr)^-1.

4.2 Electron Spectra Inversion of Dayside Auroral Arcs

Using Eqs.(2) and (4), we can determine the average energy E and total energy flux J E of precipitating electrons from the I_557.7/I_630.0 ratio and I_557.7 emission intensity. The relation between total energy flux J E and average energy E is given as

(5)

where J is total number flux of precipitating electrons ((cm²·s·sr)^-1). Fig. 9 shows average energy E, total energy flux J E and total number flux J as a function of MLT. The E variation curve is plotted with solid line in the top panel, and two dashed lines are 600 and 220 eV, respectively. Most of auroral arcs have the average larger than 600 eV. The J E variation curve is plotted in the middle panel with the value between 10¹¹ and 10¹² eV·(cm²·s·sr)^-1. The J variation curve is plotted with solid line in the bottom panel, and the dashed line is corresponding to 10⁹ (cm²·s·sr)^-1. Based on the previous analysis of the cases in different source regions, the criterion for determining the source region of dayside auroral arcs can be summarized as follows: the auroral arcs with average energy more than 600 eV were classified as BPS marked by red point; the auroral arcs with average energy less than 600 eV and location (zenith angle) of arcs less than 0 ( < 76.2 MLAT) were classified as LLBL marked by green point; the auroral arcs with average energy less than 600 eV and location (zenith angle) of arcs larger than 0 (≥76.2 MLAT) were classified as Mantle marked by blue point. From Fig. 9, the main source region is BPS in the prenoon and postnoon. The number of Mantle auroral arcs is few, located in the higher latitude at dawn and dusk, which may be associated with the polar cap auroral arcs referred by Lassen and Danielsen (1978). The source region is LLBL adjacent to magnetic noon (11-13 MLT): average electron energy is between 220 and 600 eV, total electron energy flux is larger than 10¹¹ eV·(cm²·s·sr)^-1, total number flux is about 10⁹ (cm²·s·sr)^-1, located in the lower latitude.

From the bottom panel in Fig. 9, we can see the total number flux of precipitating electrons increases as it is close to magnetic noon and the maximum appears at about 1400 MLT. It is consistent with the result that the I_557.7/I_630.0 ratio is smaller near magnetic noon for the same I_557.7 emission intensity in Figs. 2 and 3, because total number flux is inversely proportional to average electron energy (~I_557.7/I_630.0 ratio) from Eq.(5). Total number flux of precipitating electrons is related to the particle density in source region. Source region is mainly LLBL near magnetic noon with the particle density of 1~40/cm³ (Sckopke et al., 1981; Li et al., 2011). More precipitating electrons come from BPS in the prenoon and postnoon sectors with the particle density of 0.1~1/cm³ which is much less than that in LLBL (Eastman et al., 1984). Therefore, total number flux of precipitating electrons is larger near magnetic noon. The maximum value appears at 1400 MLT, because anti-parallel reconnection can occur easily in the mapping source region in the magnetosphere and thus particles can enter into magnetosphere precipitating to ionosphere from solar wind, resulting in an increase of electron number flux (Hu et al., 2014). From Fig. 9, source region at near 1400 MLT can be BPS or LLBL. Some dynamic processes have not been known thoroughly in this region.

5 CONCLUSION

We have statistically analyzed the auroral arc emissions and precipitating electron spectra and discussed the source region of auroral arcs with different spectrum characteristics, surveying over 20000 multiple-wavelength and 630.0 nm) dayside auroral arc ASIs at YRS during a 4-year period, together with the particle precipitation data measured by DMSP satellites. The conclusions of this study can be summarized as follows:

(1) The 557.7 nm emission intensity has two maximum values in the hot spot and warm spot regions, with average values of 2.2 and 2.9 kR, respectively. But there is a maximum near magnetic noon for 630.0 nm emission intensity, having an average value of 1.5 kR.

(2) The I_557.7/I_630.0 ratio of auroral arcs increases with the I_557.7 emission intensity. The I_557.7/I_630.0 ratio tends to increase and takes the values 0.1~10 when the I_557.7 emission intensity increases from 0.1 to 10 kR.

(3) The average energy of precipitating electrons is proportional to I_557.7/I_630.0 ratio. There is an expo-nential relationship between total energy flux of precipitating electrons and I_557.7 emission intensity.

(4) The typical region of dayside auroral arcs is BPS in the prenoon and postnoon sectors. It is also found that the precipitating electrons come from the region of Mantle, where the arcs are located poleward of dayside auroral oval. Source region of the precipitating electrons is identified as LLBL adjacent to magnetic noon, where the arcs are located in the lower latitude.

Precipitating electron spectra can be obtained using Eqs.(2) and (4), which can help us obtain the source region of dayside auroral arcs without DMSP data. The morphological and dynamic characteristics of dayside auroral arcs in different source regions require further study and discussion.

ACKNOWLEDGMENTS

This work was supported by the National Natural Science Foundation of China (41431072, 41274164 and 41504115), the Polar Environment Comprehensive Investigation and Assessment Programs (CHINARE2016-02-03, 2016-04-01), the Strategic Priority Research Program on Space Science, Chinese Academy of Sciences (XDA04060201). Thanks are also given for the Pudong Development of Scientific and Technology Program (Pkj2013-z01), the Open Foundation of the State Oceanic Administration Key Laboratory for Polar Science (KP201303) and the Top-Notch Young Talents Program of China. We thank UAP division in Polar Re-search Institute of China to supply the auroral image data. The DMSP data is from web site: http://sd-www.jhuapl.edu/Aurora/.