Decrease in sodium density observed during auroral particle precipitation over Tromsø, Norway

Transcription

1 GEOPHYSICAL RESEARCH LETTERS, VOL. 40, , doi:10.10/grl.50897, 13 Decrease in sodium density observed during auroral particle precipitation over Tromsø, Norway T. T. Tsuda, 1,2 S. Nozawa, 3 T. D. Kawahara, 4 T. Kawabata, 3 N. Saito, 5 S. Wada, 5 Y. Ogawa, 1 S. Oyama, 3 C. M. Hall, 6 M. Tsutsumi, 1 M. K. Ejiri, 1 S. Suzuki, 3 T. Takahashi, 3 and T. Nakamura 1 Received 13 June 13; revised 15 August 13; accepted 19 August 13; published 6 September 13. [1] Using a simultaneous and common-volume observation by a European incoherent scatter (EISCAT) VHF radar and a sodium lidar at Tromsø, Norway (69.6 ı N, 19.2 ı E), we have determined, for the first time, the effect of pure particle precipitation, excluding that of the electric field, on sodium density variations. Our observation on January 12 showed that sodium atom density decreased when there was no ion temperature enhancement (indicating a weak electric field) and the electron density increased (indicating strong particle precipitation). From the results, we have concluded that auroral particle precipitation induced sodium atom density decrease in this event. Furthermore, a discussion is provided regarding the time response of the decrease in sodium density. Citation: Tsuda, T. T., et al. (13), Decrease in sodium density observed during auroral particle precipitation over Tromsø, Norway, Geophys. Res. Lett., 40, , doi:10.10/grl Introduction [2] For the last few decades, high-latitude sodium lidar observations have been conducted due to the considerable interest in the relationship between auroral particle precipitation and neutral sodium layer variations. There are several previous studies on this issue [von Zahn et al., 1987; Nomura et al., 1987; Gu et al., 1995; Heinselman et al., 1998; Heinselman, 00]. [3] von Zahn et al. [1987] reported an explosive growth of narrow sodium layers at altitudes near 95 km via a sodium lidar observation at the Andøya rocket range in Norway (69 ı N, 16 ı E). By taking into consideration the amount of sodium, they proposed that sudden or sporadic sodium layers (SSLs) were formed via the release of upper atmospheric dust caused by energetic auroral particles; however, no observational information on auroral activity was utilized. Observational results that would support the hypothesis by von Zahn et al. [1987] were obtained during the 1993 Arctic 1 National Institute of Polar Research, Tachikawa, Japan. 2 Japan Society for the Promotion of Science, Tokyo, Japan. 3 Solar-Terrestrial Environment Laboratory, Nagoya University, Nagoya, Japan. 4 Faculty of Engineering, Shinshu University, Nagano, Japan. 5 Photonics Control Technology Team, RIKEN Center for Advanced Photonics, RIKEN, Wako, Japan. 6 Tromsø Geophysical Observatory, University of Tromsø, Tromsø, Norway. Corresponding author: T. T. Tsuda, National Institute of Polar Research, 10-3 Midori-cho, Tachikawa, Tokyo , Japan. (tsuda.takuo@ nipr.ac.jp) 13. American Geophysical Union. All Rights Reserved /13/10.10/grl Noctilucent Cloud Campaign (ANLC-93) [Gu et al., 1995]. Airborne observations with a sodium lidar and an optical camera during ANLC-93 showed that a SSL began to form near a height of 94 km amid strong auroral emissions. [4] Contrary to the ANLC-93 results, Heinselman et al. [1998] reported a decrease in sodium column density during electron density enhancements due to auroral particle precipitation derived from sodium lidar and incoherent scatter (IS) radar observations at the Sondrestrom facility in Greenland (66 ı N, 50 ı W). To reproduce this observed decrease in sodium column density, Heinselman [00] conducted a modeling study and found that neutral sodium atoms can be ionized via charge transfer reactions with aurorally enhanced ionization products (O + 2 and NO+ ). From a coordinated ground-based observation at Syowa Station in Antarctica (69 ı N, 39 ı W), Nomura et al. [1987] reported that sodium column density decreased during a cosmic noise absorption (CNA) event (indicating that D region high electron density was induced by auroral particle precipitations) as well as a large variation in geomagnetic H component (indicating strong ionospheric electric fields). [5] The conflicting results from previous authors demonstrate that the response of sodium density to auroral activity is unclear; in some cases it increased, in others it decreased. [6] In addition to auroral particle precipitation, the ionospheric electric field is also considered to be an important factor in auroral related effects. In particular, such electric fields can become more significant in the vicinity of auroral precipitating regions due to field-aligned current and its current closure system in the ionosphere [e.g., Fujii et al., 09]. Ion motions driven by the electric fields could induce vertical and horizontal transportation (even convergence and/or divergence) of sodium ions [e.g., Kirkwood and von Zahn, 1991] and could then affect the generation and/or loss of sodium atoms through their ion-molecule chemistry. This effect is also considered in the probable hypothesis of SSL formation [e.g., Cox and Plane, 1998], and thus, this effect may be quite large for decreases and increases in sodium density. The effects of auroral particle precipitation and ionospheric electric fields on sodium density are not clear. Until the effects of these processes are known, accurate interpretation of the variability of sodium density will be impossible. In this paper we present the first investigation of the effect of particle precipitation using observations to ensure the absence of an electric field effect. 2. Observational Results and Discussion [7] On January 12, an Unusual Program (UP) experiment was carried out using the European incoherent 4486

2 TSUDA ET AL.: SODIUM LAYER DURING AURORA EISCAT Ne (x EISCAT Ti (K) m ) 3 4 (a) LIDAR Ns (x10 m ) (b) (c) Height (km) Figure 1. Time-height variations of (a) electron number density (Ne), (b) ion temperature (Ti), and (c) sodium number density (Ns) from 14:00 UT on 24 January to 08:00 UT on 25 January 12. scatter (EISCAT) VHF radar at Tromsø, Norway (69ı N, 19ı E). This was in response to a prediction of the impact of a coronal mass ejection related to a M8.7 class X-ray solar flare. The colocated Tromsø sodium lidar [Tsuda et al., 11] was also operated during nighttime on January 12. The distance between the EISCAT VHF radar and the sodium lidar is less than 300 m. The field-of-view (FOV) of the EISCAT VHF radar is 1.2ı in the zonal direction and 1.7ı in the meridional direction (full-width radar beam at half power), and the divergence of the laser beam of the sodium lidar is smaller than 1 mrad. Hence, the vertical beam of the sodium lidar above 30 km in height is completely within the FOV of the EISCAT VHF radar. For example, the FOVs at 90 km are km in horizontal scale for the EISCAT VHF radar and 90 m for the sodium lidar. Thus, we successfully conducted a simultaneous and common-volume observation of the ionosphere and neutral sodium layer via the EISCAT VHF radar and sodium lidar, respectively. [8] Figure 1 shows an overview of the observations: Electron density and ion temperature from the EISCAT radar, and sodium density from the sodium lidar are shown in Figures 1a 1c, respectively. The integration time and height for the observational parameters of both instruments are 2 min and 3 km, respectively. Enhancements of electron density due to auroral particle precipitation were frequently seen at :00 :00 and 22:00 08:00 UT. For ion temperature above 1 km, which can be used as an indicator of frictional heating caused by the ionospheric electric field, small enhancements were detected several times at 16:00 04:00 UT, while no enhancements were observed during 14:00 16:00 and 04:00 08:00 UT. The data coverage of the lidar was good between 80 and 100 km. Some wavelike structures were seen in the sodium density variation. [9] Here we focus on data when the electron density enhancements reached below 100 km (i.e., around 18:40, 22:00, and 04:00 :00 UT) in order to investigate the effects of auroral precipitation on sodium density. The aurorae above Tromsø were discreet-type aurorae at approximately 18:40 and 22:00 UT and diffuse-type pulsating aurorae at 04:00 :00 UT (see Figure 2). At approximately 18:40 and 22:00 UT, the ion temperature was slightly enhanced, which indicates some electric field injection (see also Figure 3c). On the other hand, no ion temperature enhancements were observed at 04:00 :00 UT, suggesting no (or very weak) electric field injection. For the diffuse-type aurorae at 04:00 :00 UT, the motion of the pulsating aurorae patch could also be used as an indicator of the ionospheric drift speed, i.e., the ionospheric electric field [cf. Scourfield et al., 1983]. The speed of the patch motions, which was measured by an all-sky Watec imager, was smaller than m s 1 during this time period, and its corresponding electric field was smaller than 1 mv m 1 (not shown here). This result also suggests that the electric field was very weak during 04:00 :00 UT. Thus, in order to investigate the actual effect of the particle precipitation on the sodium density variations without the presence of an electric field injection, the diffuse-type aurorae at 04:00 :00 UT are considered suitable. [10] The nightly averaged sodium number density and deviation from the nightly averaged sodium density at each height are shown together with the ion temperature at 154 km in Figure 3. At 04:00 :00 UT, a decrease in sodium Figure 2. All-sky images above Tromsø on January 12, captured by a color digital camera with 15 s exposure. 4487

3 100 (a) Ns-Deviation (%) Ns-Average (x10 m ) (b) Height (km) EISCAT Ti at 154 km (K) (c) 16:00 18:00 :00 22:00 00:00 :00 04:00 06:00 UT Figure 3. (a) Deviation from averaged sodium number density at each height (Ns-Deviation), (b) the averaged sodium number density (Ns-Average), and (c) the ion temperature (blue) at 154 km and the neutral temperature (black thick line) from the NRLMSISE-00 model [Picone et al., ]. Black and gray lines overlaid on the Ns-deviation indicate the electron densities of and m 3, respectively. density of several tens of percent at km was observed (down to 60% around 100 km). This decrease seemed to correspond well with the enhanced electron density (larger than and m 3, shown by black and gray lines) in both time and height, as shown in Figure 3a. On the other hand, there were sodium density variations of several tens of percent during periods of no electron density enhancements; for example, sodium density decreased during :00 22:00 UT at km and increased for 16:00 18:00 UT at km. Because regions of this decrease/increase seemed to be propagating downward, these variations would probably be due to upward propagating atmospheric waves (with upward energy propagation) [cf. Fritts and Alexander, 03]. In the case of the sodium density decrease during 04:00 :00 UT at km (i.e., when sodium density decreased and electron density became enhanced), such a downward phase propagation was not found (or was unclear), and thus, the decrease would not be due to atmospheric waves but associated with the auroral particle precipitation. [11] In the scatterplots (with zero lag) of the observed electron density and sodium density (see Figures 4 4c), it can be slightly seen that the sodium densities at 90, 95, and 100 km tended to decrease when the electron density increased (more than m 3 ). At these heights, the sodium density had negative correlation with the electron density, but the negative peaks of cross correlation were not so high ( 9 to 0.53) for data during the entire period (see Figures 4d 4f). The low correlation would be due to large variability that may have been induced by the electric fields and the atmospheric waves. For data during the period of 03:00 :00 UT on 25 January 12 (the period of very weak electric field), the negative correlation became higher ( 0.73 to 0.90) and the time lags of the sodium density response were 0 min at 90 km, 4 min at 95 km, and 14 min at 100 km. Thus, in this event, the sodium density response due to high electron density (i.e., the auroral particle precipitation) during low ion temperature (i.e., without the electric field injection) was a decrease (not increase), and the response time of the sodium density was longer at higher altitudes. [12] As a final point, we discuss the response time of the decrease in sodium density while assuming that the decrease is mainly induced by charge transfer reactions [Heinselman, 00], because our result would then agree with the reports by Nomura et al. [1987], Heinselman et al. [1998], and Heinselman [00]. First, high ion density (i.e., high electron density) is necessary for sodium density to decrease by charge transfer reactions, but most of the ions would be rapidly removed due to the recombination process. According to a calculation based on the widely used recombination rate [e.g., Fujii et al., 09], for example, an electron density of m 3 is decreased to less than m 3 within 30 s at km. Therefore, the effective loss of the sodium atoms should be achieved within this time scale. Second, lifetimes of the sodium ions (i.e., time scale of the reformation of the sodium atoms) are 72 s at 90 km, 18 minat95km,and4hat100km[collins et al.,, Figure 3]. Thus, at 90 km, both the loss and reformation of the sodium atoms can be achieved rather quickly compared with the time resolution of the observation (2 min) in this study. This would be consistent in that the negative correlation had a peak with no lag at 90 km (i.e., there was no time delay in the response). For heights of 95 and 100 km, loss could be faster and the reformation could be slower than the time resolution of 2 min. Slower reformation can induce time delays in the sodium density response. However, the observed time delay (4 min at 95 km and 14 min at 100 km) is much shorter than the lifetime of the sodium ions (18minat95kmand4hat100km).Thisdisagreementmay be due to the horizontal advection. Horizontal wind speeds measured by the colocated meteor radar were m s

4 LIDAR Ns (x10 ) LIDAR Ns (x10 m ) LIDAR Ns (x10 m ) (c)99.7km (f)99.7km -peak: peak: (b)95.0km (e)95.0km -peak: peak: (a)90.3km EISCAT Ne (x10 m ) (d)90.3km -peak:-9 -peak: Lag (min) Figure 4. Scatterplots (with zero lag) of the observed Ne and Ns at (a) 90.3, (b) 95.0, and (c) 99.7 km, and the correlation coefficients () (derived from the cross-correlation analysis) at (d) 90.3, (e) 95.0, and (f) 99.7 km. The correlation coefficient at negative peaks (-peak) are shown at the upper right. Black and red colors indicate data during the entire period and during a period of 03:00 :00 UT on 25 January 12, respectively. at km during 03:00 :00 UT (not shown here). Therefore, the motion of the neutral sodium atoms with the 2 min resolution corresponds to a horizontal distance of km. This horizontal scale is larger than that of the observation volumes (i.e., the FOVs at 90 km are km for the EISCAT VHF radar and 90 m for the sodium lidar). Thus, the observed time response of the sodium density can be interpreted from the time scale of the charge transfer reactions. 3. Concluding Remarks [13] The relationship between auroral particle precipitations and sodium atom layer variations is still an unresolved subject, although there are several previous studies. To examine this issue, it is important to distinguish auroral particle precipitation from ionospheric electric fields for evaluation of their actual effects on sodium atom density variations. However, the separation of these two effects has never been done in previous studies. In order to overcome this issue, we have performed a simultaneous and commonvolume observation via an EISCAT radar and a sodium lidar at Tromsø, Norway, and have investigated the auroral particle precipitation effect on the sodium density variations. In the present paper, we have reported the first finding that the actual particle precipitation effect (without electric field injection) induced a decrease (not increase) in the sodium density. [14] The concluding result that auroral particle precipitation can induce a decrease in sodium density is in agreement with the reports by Heinselman et al. [1998] and Nomura et al. [1987], but in disagreement with the reports by von Zahn et al. [1987] and Gu et al. [1995]. In Heinselman et al. [1998], it was reported that sodium density decreases during periods of high electron density below 95 km. High electron density at lower altitudes (below 95 km) indicates the presence of relatively high-energy auroral particles, which tend to be related to diffuse-type aurorae (generally) without strong electric field injection. The CNA event used in Nomura et al. [1987] is an indicator of high electron density at lower altitudes (in D-region), and thus their event may be also related to diffuse-type aurorae. On the other hand, the strong auroral emissions reported in Gu et al. [1995] may indicate discreet-type aurorae, which tend to be related to strong electric field injection. Thus, the type of aurorae would be an important factor for the auroral particle energy and the related electric field injection. [15] To further examine our concluding result, we would need more events and also we should pay more attention to sodium density variations induced by wavelike structures (i.e., atmospheric waves). In fact, the calculations of the Ns-deviation and the correlation coefficients used in this study do not exclude the effects of such atmospheric waves. Taking into account the atmospheric wave effects would be important for more quantitative evaluation of the sodium density variation induced by the particle precipitation. However, because the enhanced electron density and decrease in sodium density related well in both time and height, the uncertainty, if any, on the atmospheric wave effects would be very minor in the present event. [16] Acknowledgments. The Tromsø sodium lidar project is mainly supported by Special Funds for Education and Research (Energy Transport Processes in Geospace) from MEXT, Japan, under collaboration with Nagoya University, Shinshu University, RIKEN, University of Tromsø, and 4489

5 EISCAT. We are indebted to the director and staff of EISCAT for operating the facility and supplying the data. EISCAT is an international association supported by research organizations in China (CRIPR), Finland (SA), Japan (STEL and NIPR), Norway (NFR), Sweden (VR), and the United Kingdom (NERC). The all-sky color digital camera is operated by Nagoya University, and the all-sky Watec imager is operated by NIPR, Japan. The meteor radar is operated by NIPR, Japan, and University of Tromsø. We wish to appreciate X. Chu and P. J. Espy for their valuable comments. This research is partly supported by a grant-in-aid for Nagoya University GCOE Program, QFPU from MEXT, Japan; by a grant-in-aid for Scientific Research B ( , , ) from JSPS, Japan; by a grant-in-aid for Scientific Research C ( ) from JSPS, Japan; by a grant-in-aid for JSPS Fellows (25526) from JSPS, Japan; and by a Norwegian Research Council project /V30. [] The Editor thanks two anonymous reviewers for their assistance in evaluating this paper. References Collins, S. C., et al. (), A study of the role of ion-molecule chemistry in the formation of sporadic sodium layers, J. Atmos. Sol. Terr. Phys., 64, Cox, R. M., and J. M. C. Plane (1998), An ion-molecule mechanism for the formation of neutral sporadic Na layers, J. Geophys. Res., 103, Fritts, D. C., and M. J. Alexander (03), Gravity wave dynamics and effects in the middle atmosphere, Rev. Geophys., 41(1), 1003, doi:10.19/01rg Fujii, R., Y. Iwata, S. Oyama, S. Nozawa, and Y. Ogawa (09), Relations between proton auroras, intense electric field, and ionospheric electron density depletion, J. Geophys. Res., 114, A09304, doi:10.19/ 09JA Gu, Y. Y., J. Qian, G. C. Papen, G. R. Swenson, and P. J. Espy (1995), Concurrent observations of auroral activity and a large sporadic sodium layer event during ANLC-93, Geophys. Res. Lett., 22, Heinselman, C. J., J. P. Thayer, and B. J. Watkins (1998), A highlatitude observation of sporadic sodium and sporadic E-layer formation, Geophys. Res. Lett., 25, Heinselman, C. J. (00), Auroral effects on the gas phase chemistry of meteoric sodium, J. Geophys. Res., 1, 12,181 12,192. Kirkwood, S., and U. von Zahn (1991), On the role of auroral electric fields in the formation of low altitude sporadic-e and sudden sodium layers, J. Atmos. Terr. Phys., 53, Nomura, A., T. Kano, Y. Iwasaka, H. Fukunishi, T. Hirasawa, and S. Kawaguchi (1987), Lidar observations of the mesospheric sodium layer at Syowa Station, Antarctica, Geophys. Res. Lett., 14, Picone, J. M., A. E. Hedin, D. P. Drob, and A. C. Aikin (), NRLMSISE-00 empirical model of the atmosphere: Statistical comparisons and scientific issues, J. Geophys. Res., 107(A12), 1468, doi:10.19/ja Scourfield, M. W. J., J. G. Keys, E. Nielsen, C. K. Goertz, and H. Collin (1983), Evidence for the EB drift of pulsating auroras, J. Geophys. Res., 88(A10), , doi:10.19/ja088ia10p Tsuda, T. T., et al. (11), Fine structure of sporadic sodium layer observed with a sodium lidar at Tromsø, Norway, Geophys. Res. Lett., 38, L181, doi:10.19/11gl von Zahn, U., P. von der Gathen, and G. Hansen (1987), Forced release of sodium from upper atmospheric dust particles, Geophys. Res. Lett., 14,