Phase reversal of the diurnal cycle in the ... - Wiley Online Library

Click Here

JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 115, A01305, doi:10.1029/2009JA014689, 2010

for

Full Article

Phase reversal of the diurnal cycle in the midlatitude ionosphere Huixin Liu,1 Smitha V. Thampi,1 and Mamoru Yamamoto1 Received 28 July 2009; revised 4 September 2009; accepted 10 September 2009; published 30 January 2010.

[1] The typical diurnal cycle of the midlatitude F region electron density consists of a midday maximum and a midnight minimum. However, a phase reversal of this diurnal cycle has been found to occur in three distinct regions on the globe. They are the East Asian (EA) region centered around (53°N, 150°E), the Northern Atlantic (NA) region centered around (45°N, 50°W) and the South Pacific (SP) region centered around (60°S, 110°W). The intensively reported Weddell Sea Anomaly falls inside the SP region. The phase reversal occurs during March–August in EA and NA regions, and between August and March in SP region, being most prominent in local summer. Furthermore, this diurnal anomaly is more pronounced at solar minimum than at solar maximum, and more pronounced in SP region than in NA and EA regions, in terms of larger diurnal magnitude and more months it lasts in a year. It is emphasized that the diurnal anomaly consists of not only a nighttime enhancement, but also a concurrent noontime depletion. Hence, midlatitude summer nighttime enhancements reported in previous studies are just part of this reversed diurnal cycle. The cause for the phase reversal involves several interplaying physical processes. Among these, the neutral wind combined with the geomagnetic field configuration plays a pivotal role. It generates a one‐wave longitudinal pattern at southern middle latitudes and a two‐wave pattern at northern middle latitudes, whose wave peaks correspond to the center of the SP, EA, and NA regions, respectively. The seasonal variation of neutral winds and downward motion of the ionization induced by thermal contraction of the ionosphere at sunset may largely control the occurring local time of the nighttime density enhancement and how long it persists in different months. The phase reversal occurs as a result of close ion‐neutral coupling. It is further noted that winter anomaly in the EA, NA, and SP regions is very weak or missing. Citation: Liu, H., S. V. Thampi, and M. Yamamoto (2010), Phase reversal of the diurnal cycle in the midlatitude ionosphere, J. Geophys. Res., 115, A01305, doi:10.1029/2009JA014689.

1. Introduction [2] The ionosphere is the ionized part of the Earth’s upper atmosphere, which forms by the solar EUV/UV radiation via photoionization. The amount of ionization production depends on the solar zenith angle and radiation intensity. Consequently, it experiences daily and seasonal variations, also effects from the solar rotation and solar cycle. The electron density (Ne) in the E and lower F region ionosphere is proportional to the square root of the production rate and hence will experience similar variations. Such a layer is called a Chapman layer, with mean diurnal cycle consists of a midday density maximum and a nighttime density minimum [Kelley, 1989]. Historically, any departure from this simple solar‐controlled behavior has been called an “anomaly.” The major midlatitude F2 layer anomalies are the win1 Research Institute for Sustainable Humanosphere, Kyoto University, Kyoto, Japan.

Copyright 2010 by the American Geophysical Union. 0148‐0227/10/2009JA014689$09.00

ter or seasonal anomaly, semiannual anomaly and annual anomaly. These anomalies have been long‐recognized, and relatively well understood owing to extensive and intensive studies [Torr and Torr, 1973; Rishbeth, 1998; Rishbeth et al., 2000; L. Liu et al., 2009, and references therein]. [3] Deviations from the mean diurnal cycle have also been reported before. For instance, Bellchambers and Piggott [1958] and Dungey [1961] found unusual local summer nighttime enhancements in foF2 at Faraday (65.2°S, 64.6°W), which they named as the “Weddell Sea Anomaly” (WSA). Similar phenomenon was found to occur at Port Lockroy (65.8°S, 63.5°W) [Rastogi, 1960] and also in the northern hemisphere at Millstone Hill (42.6°N, 71.5°W) [Evans, 1965]. Global observations from the CHAMP satellite showed unusual diurnal variations of Ne in summer midlatitudes in both hemispheres [Liu et al., 2007b]. Recently, detailed climatology of the WSA has been reported based on modern satellite observations [Horvath and Lovell, 2009; Lin et al., 2009; Jee et al., 2009; He et al., 2009, and references therein]. A WSA‐like feature in the northern hemisphere was also noticed by Lin et al. [2009]. This fea-

A01305

1 of 10

A01305

LIU ET AL.: PHASE REVERSAL OF THE DIURNAL CYCLE

A01305

Figure 1. Geographical distribution of Ne at 400 km at night (2200 LT), day (1200 LT), and the difference between them for (left) December solstice and (right) June solstice. Note that different color scales are used to clearly show the relevant patterns in each plot. ture has been confirmed by Thampi et al. [2009] using the tomographic observations near 135°E longitude. Together with the WSA, such features are generally named as the midlatitude summer nighttime anomaly (MSNA) [Thampi et al., 2009]. These observations are either from the Ionosonde, Total Electron Content (TEC), tomography or radio occultation (both need an inversion procedure). To compliment such observations, we use in situ measurements from the CHAMP satellite to examine this phenomenon. Furthermore, we investigate it from a different perspective. Instead of focusing on the nighttime enhancement as done in many previous studies, we view it as part of the diurnal cycle and examine the daytime and nighttime behavior as a whole.

2. Data [4] CHAMP is in a near‐circular orbit with an inclination of 87.3° and an initial height of about 450 km at launch in July 2000. Its orbital plane precesses through all local times every four months and through all longitudes at a fixed local time every 24 h. In situ measurements of Ne from the Planar Langmiur Probe (PLP) on board CHAMP satellite are uti-

lized in this study. These measurements are of 15 s resolution, hence giving a large sampling set with global coverage. [5] The chosen data period spans from January 2001 to December 2006. Only data under quiet geomagnetic conditions with Kp ≤ 3 are used in the following analysis to limit effects from geomagnetic disturbances. The data are subdivided into two groups, with one representing solar maximum (SMAX) (F10.7 ≈ 160) and one representing solar minimum (SMIN) (F10.7 ≈ 90). Furthermore, since the CHAMP orbit has decayed from 456 km height to about 350 km during these 6 years, a normalization of the data to a common altitude of 400 km has been applied to possibly exclude variations induced by the orbit decay. The normalization is done using the IRI2000 model [Bilitza, 2003] Þ , where h denotes as following: NC (400 km) = NC (h) NI ðN400km I ðhÞ CHAMP’s orbital height (h), NC and NI denote Ne observed by CHAMP and predicted by IRI2000 model at the same location, respectively. Although this procedure may introduce some uncertainties from the model to the data, for the analysis of the diurnal cycle and its seasonal and solar activity dependence, it is highly necessary to avoid density variations due to large altitude changes. Such normalization

2 of 10

A01305


procedure has been successfully applied in previous studies [Liu et al., 2007a]. Ionosonde data at Wakkanai (45°N, 142°E) are also employed for comparison.

3. Results 3.1. Distribution of the Anomaly Diurnal Cycle Regions [6] Geographical distributions of Ne at 400 km in June and December are depicted in Figure 1 to illustrate the diurnal anomaly regions at a moderate solar flux level (F10.7 ≈ 130). Nighttime Ne patterns represented by that at 2200 LT are shown in Figure 1 (top), while noontime patterns around 1200 LT are shown in Figure 1 (middle). Density differences between day and night (DNe = Nenight − Neday) are given in Figure 1 (bottom). The absolute difference is preferred to the percentage one to avoid possible exaggeration by small Ne values. [7] At both 1200 LT and 2200 LT, Ne shows familiar Equatorial Ionization Anomaly (EIA) at equatorial and low‐latitude regions. A wave‐3/wave‐4 longitudinal modulation is also discernible at 1200 LT/2200 LT. Here we are interested in the density difference between day and night. As seen in Figure 1 (bottom), DNe is mostly negative, indicating higher Ne at day than at night. However, there are also distinct regions with positive DNe, signaling a diurnal anomaly with higher density at night than at day. In December, one such anomaly region exists near (60°S, 110°W) in the South Pacific (SP), with a large spatial extension between 160°E–30°W and 30°S–85°S. In its center region, Ne at night is more than triple of that around noon. A tilt of this region following the excursion of the dip equator is clearly visible, which may give us a hint on its connection to the geomagnetic field. In June, two anomaly regions appear in the northern hemisphere, with one centering around (50°N, 150°E) in East Asia sector (EA), and the other centering at (45°N, 50°W) in the North Atlantic (NA). Each has a spatial coverage about 8000 km in the east‐west direction and ∼3000 km in the north‐south direction, hence being about one‐third of the SP area. In addition, the nighttime density is about double of the daytime one. Thus, the diurnal anomaly appears to be most pronounced in the SP region, in terms of either the amplitude of the diurnal cycle (defined as the ratio of the nighttime maximum to noontime minimum) or the spatial coverage. 3.2. Month‐to‐Month Variation of the Diurnal Cycle [8] After seeing “where” the diurnal anomaly occurs, we now investigate “when” it occurs in terms of season. Taking locations near the center of these anomaly regions, we investigate the month‐to‐month variation of the diurnal cycle. Results are shown in Figures 2 and 3 for solar minimum and solar maximum, respectively. In Figures 2 and 3, Ne at 400 km is presented in the month versus local time frame for SP and EA regions. Line plots for solstices in corresponding regions are also given in Figures 2 and 3 (right) to show the diurnal features more clearly. The NA region is not shown since it exhibits nearly identical behavior to the EA region. The NmF2 measured by an Ionosonde at Wakkanai in 2008 (SMIN) and 2001 (SMAX) are presented as well for comparison and validation purposes. [9] Let us first examine the solar minimum condition shown in Figure 2. In SP region, Ne maximizes near 1400 LT in

A01305

June (local winter), gradually decreasing toward both dawn and dusk side as shown by the red line in Figure 2 (top right). This is the normal diurnal cycle of Ne at middle latitudes, which is largely controlled by the local time variation of the solar zenith angle. Conversely in December (local summer), Ne maximizes near midnight and minimizes around noon (black line in Figure 2, top right). Therefore, a reversal occurs in December in terms of the phase of the diurnal cycle. From Figure 2 (top left), we see that the phase reversal occurs not only in December, but persists throughout August–March. Furthermore, the nighttime maximum seems to shift toward later local times from August toward December, then back to earlier local times from December toward March. [10] In East Asia region (Figure 2, middle), the anomalous diurnal cycle occurs during March–August, with Ne maximizes at night, and minimizes around noon. The phase reversal occurs most prominently in June (local summer). Similar to the SP region, the maximum Ne has a tendency to shift from 1900 LT in April to 2200 LT in June, then shift back toward earlier local times in later months, e.g., 2000 LT in August. We notice that high nighttime Ne lasts for only about 2 h and is mainly confined to premidnight sector. In comparison, high values of Ne last throughout the night and till even 0400 LT in the SP region. The NmF2 observed in 2008 by an Ionosonde in Wakkanai, which is inside the EA region, exhibits a noontime minimum and a nighttime maximum through out April–August. This corroborates the CHAMP observation at 400 km, but with a smaller nighttime‐to‐noontime ratio (2.4 for Ne at 400 km, and 1.4 for NmF2). [11] At solar maximum (Figure 3), the month‐to‐month variation of the diurnal cycle basically resembles that at solar minimum, but with two differences. First, the amplitude is smaller. For instance, the summer nighttime‐to‐noontime ratio is about 2.7 in comparison to 5 in SP region (black curves in Figures 2 and 3, top right) and about 2 in comparison to 2.4 in the EA region. Second, the period of the anomaly diurnal cycle is shorter. This is especially so for the EA region, where the diurnal anomaly occurs during May–July in comparison to March–August at solar minimum. Thus, the diurnal anomaly appears to be more prominent at solar minimum in terms of either the amplitude or the range of months it lasts. [12] Note that in these diurnal anomaly regions, Ne is higher in local summer than in local winter at all local times at solar minimum. In comparison, the noontime NmF2 is higher in winter than in summer, showing the winter anomaly. Thus, these regions of diurnal anomaly are devoid of winter anomaly when seen at one fixed altitude. This feature persists even at solar maximum in the SP region. Furthermore, these anomaly regions are also regions with very weak semiannual variation of the noontime Ne as compared to normal regions, e.g., at (50°N, 60°E) as illustrated in Figure 4. Such features relevant to the winter and semiannual anomalies are interesting, but will not be discussed further.

4. Discussion [13] The above analysis has revealed three distinct regions on the globe with reversed diurnal cycle of the electron density. This reversed diurnal cycle exhibits a noontime density minimum and a concurrent nighttime density maximum,

3 of 10

A01305


A01305

Figure 2. Month‐to‐month variation of diurnal cycle of Ne at 400 km in the South Pacific and East Asia regions, along with NmF2 observed by an Ionosonde at Wakkanai. The solar flux level is F10.7 ≈ 90. Note that a phase reversal of the diurnal cycle occurs around local equinoxes and summer, while the electron density maximizes at night instead of at noon. with the nighttime‐to‐noontime ratio reaching as high as five in local summers. The occurrence time of the nighttime maximum has a tendency to shift to later times from spring toward summer, then shift to earlier times from summer toward autumn.

4.1. Comparison With Previous Results 4.1.1. Anomaly Regions [14] The region of Weddell Sea Anomaly, which has been intensively reported [e.g., Bellchambers and Piggott, 1958; Horvath, 2006; Jee et al., 2009; Lin et al., 2009; He et al.,

4 of 10

A01305


A01305

Figure 3. Same as Figure 2, but for solar maximum conditions with F10.7 ≈ 160. 2009], falls inside the SP region revealed in Figure 1. Millstone Hill reported by Evans [1965] is inside the NA region, while the 136°E longitude region shown by Thampi et al. [2009] is inside the EA region. Thus, the anomaly regions revealed in Figure 1 of this study unify previously reported sites into one global map, with a full view of their spatial coverage. It is important to note that the EA and NA regions are not connected to each other. For instance, no diurnal

anomaly occurs at (50°N, 60°E) at either SMIN or SMAX as shown in Figure 4. Furthermore, we would like to point out that the MSNA (including the WSA) reported in previous studies and the diurnal anomaly reported here are actually the same phenomenon, but seen from different perspectives. By saying “diurnal anomaly,” we synthesize the daytime depletion and nighttime enhancement into the perspective of a diurnal cycle, while the name MSNA emphasizes

5 of 10

A01305


A01305

Figure 4. Month‐to‐month variation of diurnal cycle of Ne at 400 km at 50°N, 60°E. (top) Solar minimum; (bottom) solar maximum. Due to the large solar cycle variation of Ne, different color scales are used to clearly show the patterns. Note that no phase reversal of the diurnal cycle occurs at this location. the enhancement as a nighttime feature. To some extent, these two terms can be used interchangeably to refer to the same phenomenon. 4.1.2. Month‐to‐Month Variation [15] In spite of many studies on the summer nighttime Ne enhancement, no month‐to‐month variation as detailed as we obtain here has ever been reported before for any of these three regions, possibly due to limited number of observations. A rough seasonal variation in the WSA region was investigated by Jee et al. [2009] by binning their data into four seasons. They found that the nighttime enhancement of TEC occurs only in summer (November–February) at solar minimum, but also at equinoxes (March, April, September, October) at solar maximum. Meanwhile, our results show that WSA falls inside the SP region, where nighttime Ne at 400 km altitude increases substantially throughout August–March, nearly independent of solar activity. The cause for this difference could be that TEC comes mainly from the contribution of NmF2, which has a different behavior from Ne at 400 km as shown in Figures 2 and 3, and also will be discussed later. The way of analysis may also somewhat contribute to the difference. We quantitatively examine the phenomenon by making a subtraction between the day and night values instead of making a visual comparison of the daytime and nighttime pattern. Visual examination may lead to inaccurate estimations when the electron density drops to very low levels at solar minimum. [16] In the northern hemisphere, our results show that the diurnal anomaly occurs during March–August at solar minimum, and May–July at solar maximum. The amplitude of the anomaly diurnal cycle is larger at solar minimum in

terms of the nighttime‐to‐noontime ratio of Ne. This is in good agreement with the incoherent scatter radar observations at Millstone Hill [Evans, 1965]. With limited data, Evans [1965] reported that NmF2 increases at night around equinoxes and summer, with larger magnitude at solar minimum than at solar maximum. Therefore, we may state that the diurnal anomaly or MSNA is more pronounced at solar minimum than at solar maximum when looking either at a fixed altitude or NmF2. 4.2. Possible Mechanisms [17] A careful examination of the diurnal anomaly regions in Figure 1 (bottom) reveals a tilt, which is best seen in the SP and NA regions. It nicely follows the excursion of the dip equator, hence may hint on its connection to the geomagnetic field. Such a connection was actually already pointed out by Rastogi [1960]. He explained that the extremely sparse magnetic field lines in the WSA region could affect the plasma distribution via plasma transport along the field lines. Similar idea can be found in recent papers as well [Horvath, 2006; Jee et al., 2009], where these authors have advanced it by combining it with the neutral wind variation. Although these explanations were made mainly for the nighttime enhancement only, we show in the following that the same mechanism can be invoked to explain the daytime depletion and nighttime enhancement as a whole, hence the anomaly diurnal cycle. [18] As extensively discussed in the excellent review by Titheridge [1995], the neutral wind plays a predominant role in the variation of F region parameters at middle latitude by lifting up or lowering the F region ionization. Its actual

6 of 10

A01305


A01305

Figure 5. Effective wind produced by a zonal and meridional wind with speed of 100 m/s at 60°S and 50°N. The solid lines represent the nighttime situation when the wind directs eastward and equatorward, while the dashed lines represent the daytime situation when the wind direction reverses. Note that the effective wind at SP is 1.33 times of that at NA and EA. effect, We f f, depends on the configuration of the magnetic field. Let U and V be the zonal (eastward positive) and meridional (equatorward positive) wind components, D and I be the magnetic declination and inclination angles, We f f can be expressed as We f f ¼ ðV cos D U sin DÞ cos I sin I:

ð1Þ

[19] Here, plus and minus signs apply to the southern and northern hemisphere, respectively. Thus, even a longitudinally uniform neutral wind can produce significant longitudinal variation in the plasma density due to the variation of D and I. This effect is illustrated in Figure 5. Here, by assuming a constant eastward and equatorward wind at night (U = 100 m s−1, V = 100 m s−1), and westward and poleward wind at day (U = −100 m s−1, V = −100 m s−1), we estimated the longitudinal variation of We f f at 50°N and 60°S, which correspond to the center latitudes of the EA and SP regions. [20] Several features clearly stand out from this simple illustration. First, the nighttime (solid lines) upward wind maximizes around 90°W at 60°S, and at about 150°E and 40°W at 50°N. During daytime (dashed lines), the opposite wind produces the strongest downward wind at the same location. Consequently, the ionosphere will be lifted up to regions of lower recombination rate at night and will be

pushed down to regions of high recombination rate at day. This potentially leads to Ne enhancement at night (in combination with extra ionization) and depletion at noon, hence making a reversal of the normal diurnal cycle most probable at these three longitudinal sectors. We see that these three longitudinal sectors correspond nicely to the center of the three diurnal anomaly (MSNA) regions shown in Figure 1. Thus, the combined effects of neutral wind and the geomagnetic configuration can potentially explain both the daytime depletion and the nighttime enhancement, hence the phase reversal of the normal diurnal cycle. Note that this combined effect produces a one‐wave longitudinal pattern at 60°S, but a two‐wave pattern at 50°N. The peaks of these wave patterns corresponds almost perfectly to the center of the SP, EA and NA regions, respectively. Also note that owing to the unusually large declination angle near the South American anomaly, the same horizontal wind produces an effective wind (We f f) in the SP region about 1.33 times of that in the EA or NA region. This may partly explain why the diurnal anomaly is more pronounced in the SP region than in other regions. Photoionization before sunset, is a necessary factor to combine with the uplift of the ionosphere to create nighttime enhancement as discussed in detail by He et al. [2009]. [21] The above mechanism based on the neutral wind effects in the geomagnetic frame explains nicely the longitu-

7 of 10

A01305


A01305

Figure 6. Local time gradients of the electron temperature observed by CHAMP at solar minimum in different months. dinal variation of the anomaly region. However, uncertainties remain when it comes to explain why the anomaly does not occur in winter. As discussed by He et al. [2009], the relative timing of the sunset and the equatorward turning of the meridional wind is an important controlling factor. Meridional wind at several northern stations have been reported to be turn equatorward later and also weaker in winter than in summer [Kawamura et al., 2000], which may lead to the absence of phenomenon in winter in combination with an early sunset. But there is very little wind observations at other sites to generally confirm this as pointed out by Jee et al. [2009]. [22] In addition to this neutral wind process mentioned above, there is another mechanism, which can significantly contribute to the seasonal variation. This mechanism is based on the thermal contraction of the ionosphere as put forward by Evans [1965]. At sunset, the electron temperature (Te) experiences a rapid fall, which can induce fast downward motion of the ionization from above hmF2 and from the plasmasphere, as a consequence of reduced plasma scale height (which is inversely proportional to the plasma temperature). This downward motion then leads to a pile up of plasma near and above hmF2. A rapid change in Te is most likely at low geographic latitudes in summer when sunset is rapid due to the greater variation in the solar zenith angle on a summer day than on a winter day. On the other hand, downward motion is most efficient at high latitudes with high geomagnetic field inclination [Ratcliffe and Weekes, 1960]. Therefore, this mechanism is expected to be most effective at middle latitudes in summer, where these conflicting requirements are best met. We have examined the simultaneous observations of Te from CHAMP, and its local time gradient is shown in Figure 6. We see that Te falls very rapidly at sunset as indicated by the dark blue belt. The drop in summer is particularly sharp, reaching 500 K/h, which is twice as large as that in winter. Another contribu-

tory factor is the difference in the amount of ionization stored above hmF2. The scale height in summer is much larger than in winter, so a larger reservoir of ionization is available to participate in the downward motion. This argument is supported by observations, which show that the daytime ratio of the number of Ne above hmF2 to the number below hmF2 is about 2.5 in winter [Evans and Taylor, 1961], and about 5 in summer [Taylor, 1964]. Thus, the thermal contraction of the ionosphere can significantly contribute to the seasonal variation of the diurnal cycle via more efficient maintaining or even increasing Ne in the upper F region after sunset in summer. Since this process commences at sunset, we can expect the maximum Ne to occur latest in summer and moves to earlier local time toward equinoxes. Such a shift in local time roughly fits with results in Figures 2 and 3, hence offering supporting evidence for this mechanism. Note that the thermal contraction discussed here is the contraction of the ionosphere, and should not be confused with the thermospheric contraction which has little effect on NmF2 [Oliver et al., 2008]. Furthermore, the downward motion of ionization discussed here is along the geomagnetic field line, not in the vertical direction. Therefore, vertical TEC observations, which intersects many flux tubes, are expected to see an increase as well. And such TEC increases have been observed [Horvath, 2006; Jee et al., 2009]. [23] Combination of the neutral wind effect in the geomagnetic frame, the photoionization, and the thermal contraction of the ionosphere at sunset can therefore explain most features we observed. [24] One expectation from these mechanisms is that the diurnal anomaly/MSNA should be more prominent at altitudes near and above hmF2. This is because, chemical recombination occurs fast below hmF2, hence forbids any significant effects of the neutral wind or plasma motion. Above hmF2, the mean lifetime of the ions exceeds 1 h,

8 of 10

A01305


hence becomes very dependent on dynamical effects caused by neutral winds or plasma motion. In addition, since the whole ionosphere is moved up by nighttime equatorward wind and pushed down by daytime poleward wind, hmF2 is higher at night than at day as shown by He et al. [2009]. This makes a fixed altitude above hmF2, say 400 km, being relatively closer to hmF2 at night than at day. These physical and geometrical effects combine together to explain why the diurnal anomaly or MSNA is more prominently seen in Ne at 400 km than in NmF2 (see Figure 2). In the same line, the diurnal anomaly in TEC is less pronounced than in Ne at a fixed altitude above hmF2 since NmF2 is the main contributor to TEC [Lin et al., 2009]. [25] Finally, we would like to add a word on the possible relation of these anomaly regions to the wave‐4 or wave‐3 patterns of the EIA [Immel et al., 2006; Wan et al., 2008; Liu and Watanabe, 2008; H. Liu et al., 2009]. Figure 1 shows that EIA exhibits a wave‐4 pattern around 2200 LT at both solstices, while more like a wave‐3 pattern around 1200 LT. The anomaly regions shown in Figure 1 (bottom) are of larger scale size than those wave‐4 or wave‐3 crests. They appear to be rather in the regions where the trough of the wave‐4/wave‐3 occurs, if any relation between them can be claimed. Furthermore, given that meridional wind is poleward at day and equatorward at night, transport of plasma by wind from the EIA region to the diurnal anomaly regions is only probable at day.

5. Summary [26] Analysis of 6 years of in situ electron density measurements at 400 km from the CHAMP satellite has revealed three regions on the globe with diurnal anomaly of Ne. They are the named the EA, NA and SP regions, centering around (50°N, 150°E), (45°N, 50°W) and (60°S, 110°W), respectively. The EA and NA each has a spatial coverage about 8000 km in the east‐west direction, and about 3000 km in the north‐south direction. The SP region extends over 85°S, and covers an area nearly three times of that of the EA or NA region. These regions correspond to regions of MSNA. In these regions, we find (1) a diurnal anomaly with maximum Ne at night and a minimum Ne at noon occurs throughout several months during local summer and equinoxes; (2) the diurnal anomaly is seen to be most prominent in the SP region where the declination angle is largest, with peak‐to‐trough diurnal amplitude being twice as large as that in the EA and NA regions; (3) the nighttime density enhancement is confined to premidnight sector in the EA and NA region, but persists to postmidnight sector in the SP region; (4) this diurnal anomaly becomes less pronounced at solar maximum, in terms of smaller magnitude and shorter period it lasts during a year. It is emphasized that the diurnal anomaly consists of not only a nighttime enhancement, but also a concurrent noontime depletion. Hence, midlatitude summer nighttime enhancements reported in previous studies are just part of this reversed diurnal cycle. [27] In this paper, we propose a working model that may reasonably explain the formation of the diurnal anomaly and MSNA. It is the combination of neutral wind effect in the geomagnetic frame, photoionization and the thermal contraction of the ionosphere at sunset. Among these, the regulation of neutral wind effect by the magnetic field configuration,

A01305

particular by the magnetic declination, contributes largely to the longitudinal distribution of the diurnal anomaly regions. Seasonal variations of the diurnal anomaly can be ascribed to variations of the wind and the downward motion of the ionization induced by thermal contraction of the ionosphere at sunset. The relative importance of these processes under different seasonal and solar conditions should be clarified with coupled thermosphere‐ionosphere models to improve our understanding on the diurnal anomaly and the MSNA. [28] Acknowledgments. We thank Claudia Stolle for running the IRI model. The work of H.L. and S.T. is supported by the JSPS foundation. The CHAMP mission is supported by the German Aerospace Center (DRL) in operation and by the Federal Ministry of Education and Research (BMBF) in data processing. [29] Zuyin Pu thanks the reviewer for his/her assistance in evaluating this paper.

References Bellchambers, W. H., and W. R. Piggott (1958), Ionospheric measurements made at Halley Bay, Nature, 182, 1596–1597, doi:10.1038/1821596a0. Bilitza, D. (2003), International reference ionosphere 2000: Examples of improvements and new features, Adv. Space Res., 31, 757–767. Dungey, J. W. (1961), Interplanetary magnetic field and the auroral zones, Phys. Rev. Lett., 6, 47–48, doi:10.1103/PhysRevLett.6.47. Evans, J. V. (1965), Cause of the midlatitude evening increase in foF2, J. Geophys. Res., 70, 1175–1185. Evans, J. V., and G. N. Taylor (1961), The electron content of the ionosphere in winter, Proc. R. Soc. London, 263, 189–211. He, M., L. Liu, W. Wan, B. Ning, B. Zhao, J. Wen, X. Yue, and H. Le (2009), A study of the Weddell Sea Anomaly observed by FORMOSAT‐ 3/COSMIC, J. Geophys. Res., 114, A12309, doi:10.1029/2009JA014175. Horvath, I. (2006), A total electron content space weather study of the nighttime Weddell Sea Anomaly of 1996/1997 southern summer with TOPEX/Poseidon radar altimetry, J. Geophys. Res., 111, A12317, doi:10.1029/2006JA011679. Horvath, I., and B. C. Lovell (2009), Investigating the relationships among the South Atlantic Magnetic Anomaly, southern nighttime midlatitude trough, and nighttime Weddell Sea Anomaly during southern summer, J. Geophys. Res., 114, A02306, doi:10.1029/2008JA013719. Immel, T. J., E. Sagawa, S. L. England, S. B. Henderson, M. E. Hagan, S. B. Mende, H. U. Frey, C. M. Swenson, and L. J. Paxton (2006), Control of equatorial ionospheric morphology by atmospheric tides, Geophys. Res. Lett., 33, L15108, doi:10.1029/2006GL026161. Jee, G., A. G. Burns, Y.‐H. Kim, and W. Wang (2009), Seasonal and solar activity variations of the weddell sea anomaly observed in the topex total electron content measurements, J. Geophys. Res., 114, A04307, doi:10.1029/2008JA013801. Kawamura, S., Y. Otsuka, S.‐R. Zhang, S. Fukao, and W. L. Oliver (2000), A climatology of middle and upper atmosphere radar observations of thermospheric winds, J. Geophys. Res., 105, 12,777–12,788, doi:10.1029/2000JA900013. Kelley, M. C. (1989), The Earth’s Ionosphere, Academic, New York. Lin, C. H., J. Y. Liu, C. Z. Cheng, C. H. Chen, C. H. Liu, W. Wang, A. G. Burns, and J. Lei (2009), Three‐dimensional ionospheric electron density structure of the Weddell Sea Anomaly, J. Geophys. Res., 114, A02312, doi:10.1029/2008JA013455. Liu, H., and S. Watanabe (2008), Seasonal variation of the longitudinal structure of the equatorial ionosphere: Does it reflect tidal influences from below?, J. Geophys. Res., 113, A08315, doi:10.1029/ 2008JA013027. Liu, H., C. Stolle, M. Förster, and S. Watanabe (2007a), Solar activity dependence of the electron density at 400 km at equatorial and low latitudes observed by CHAMP, J. Geophys. Res., 112, A11311, doi:10.1029/ 2007JA012616. Liu, H., C. Stolle, S. Watanabe, T. Abe, M. Rother, and D. L. Cooke (2007b), Evaluation of the IRI model using CHAMP observations in polar and equatorial regions, Adv. Space Res., 39, 904–909, doi:10.1016/ j.asr.2006.08.006. Liu, H., M. Yamamoto, and H. Lühr (2009), Wave‐4 pattern of the equatorial mass density anomaly: A thermospheric signature of tropical deep convection, Geophys. Res. Lett., 36, L18104, doi:10.1029/ 2009GL039865.

9 of 10

A01305


Liu, L., W. Wan, B. Ning, and M.‐L. Zhang (2009), Climatology of the mean total electron content derived from GPS global ionospheric maps, J. Geophys. Res., 114, A06308, doi:10.1029/2009JA014244. Oliver, W. L., S. Kawamura, and S. Fukao (2008), The causes of midlatitude F layer behavior, J. Geophys. Res., 113, A08310, doi:10.1029/ 2007JA012590. Rastogi, R. G. (1960), Abnormal features of the region of the ionosphere at some southern high‐latitude stations, J. Geophys. Res., 65, 585–592. Ratcliffe, J. A., and K. Weekes (1960), The ionosphere, in Physics of the Upper Atmosphere, edited by J. A. Ratcliffe, p. 377, Academic, New York. Rishbeth, H. (1998), How the thermospheric circulation affects the ionospheric F2‐layer, J. Atmos. Sol. Terr. Phys., 60, 1385–1402. Rishbeth, H., L. Zou, I. C. F. Muller‐Wodarg, T. J. Fuller‐Rowell, G. H. Millward, R. J. Moffett, D. W. Idenden, and A. D. Aylward (2000), Annual and semiannual variations in the ionospheric F2‐layer: II. Physical discussion, Ann. Geophys., 18, 945–956.

A01305

Taylor, G. N. (1964), The electron content of the ionosphere at middle latitudes in summer, Proc. R. Soc. London, 279, 497–509. Thampi, S., C. H. Lin, H. Liu, and M. Yamamoto (2009), First tomographic observations of the midlatitudes summer night anomaly (MSNA) over Japan, J. Geophys. Res., 114, A10318, doi:10.1029/2009JA014439. Titheridge, J. E. (1995), Winds in the ionosphere—A review, J. Atmos. Terr. Phys., 57, 1681–1714. Torr, M. R., and D. G. Torr (1973), The seasonal behaviour of the F2‐layer of the ionosphere, J. Atmos. Terr. Phys., 35, 2237–2251. Wan, W., L. Liu, X. Pi, M.‐L. Zhang, B. Ning, J. Xiong, and F. Ding (2008), Wave number‐4 patterns of the total electron content over the low latitude ionosphere, Geophys. Res. Lett., 35, L12104, doi:10.1029/ 2008GL033755. H. Liu, S. V. Thampi, and M. Yamamoto, Research Institute for Sustainable Humanosphere, Kyoto University, Kyoto 6110011, Japan. ([email protected]‐u.ac.jp)

10 of 10