Comparison of middle- and low-latitude sodium layer from a ground-based lidar network, the Odin satellite, and WACCM-Na model

. The ground-based measurements obtained from a lidar network and the six-year OSIRIS limb-scanning radiance measurements made by the Odin satellite are used to study the climatology of the middle- and low-latitude sodium (Na) layer. Up to January 2021, four Na resonance fluorescence lidars at Beijing (40.2 ◦ N, 116.2 ◦ E), Hefei (31.8 ◦ N, 117.3 ◦ E), Wuhan (30.5 ◦ N, 114.4 ◦ E), and Haikou (19.5 ◦ N, 109.1 ◦ E) collected vertical profiles of Na density for a total of 2,136 nights (19,587 h). These large datasets provide routine long-term measurements of the Na layer with exceptionally high temporal and vertical 5 resolution. The lidar measurements are particularly useful for filling in OSIRIS data gaps since the OSIRIS measurements were not made during the dark winter months because they utilise the solar-pumped resonance fluorescence from Na atoms. The observations of Na layers from the ground-based lidars and the satellite are comprehensively compared with a global model of meteoric Na in the atmosphere (WACCM-Na). The lidars present a unique test of OSIRIS and WACCM, because they cover the latitude range along 120 ◦ E longitude in an unusual geographic location with significant gravity wave generation. In 10 general, good agreement is found between lidar observations, satellite measurements, and WACCM simulations. Whereas the Na number density from OSIRIS is slightly larger than that from the Na lidars at the four stations within one standard deviation of the OSIRIS monthly average, particularly in autumn and early winter arising from significant uncertainties in Na density retrieved from much less satellite radiance measurements. WACCM underestimates the seasonal variability of the Na layer at the lower latitude lidar stations (Wuhan and Haikou). discrepancy suggests the seasonal variability of vertical constituent transport modeled in WACCM is underestimated because of the gravity wave spectrum is captured in layers observed at Wuhan and Haikou lidar stations. This discrepancy suggests the seasonal variability of vertical transport of constituents is underestimated in WACCM because much of gravity waves is not resolved in model. The close relationship between the ionospheric Es layer, the Na layer, and SSLs indicates that it is important to consider the dynamical and electrodynamical processes of metallic ions in the lower E region of the ionosphere coupling with the Na layer into a global atmospheric model of metals.

decade (Wang, 2010). The lidars range from low-to mid-latitudes roughly along 120 • E longitude in China. These lidars are in an unusual geographic location with significant gravity wave generation (Hocke et al., 2016;Zeng et al., 2017). By the year 2021, a total of 2,136 nights (19,587 h) of vertical profiles of Na density were obtained. The lidar observations provide an important test of Odin satellite measurements and WACCM simulations, allowing us to investigate the impact of waves and turbulence fluctuations on the WACCM predictions of Na density. It should be noted that the original ground-truthing of Odin 70 measurements involved Na lidar measurements at Fort Collins, Colorado which is at a latitude of 41 • N (Gumbel et al., 2007).
The present paper is therefore a study of the latitudinal and seasonal variations of the Na layer at mid-and low-latitudes, using a combination of the Odin satellite measurements, the long-term ground-based measurements obtained from the lidar network, and the WACCM-Na model. The long-term Na lidar data are compared to the sun-synchronous satellite measurements at descending and ascending nodes and simultaneous WACCM simulations. Sect. 2 describes the instruments and datasets. In 75 Sect. 3, the global climatology of the Na layer from OSIRIS spectrometer with a high spatial resolution is presented for comparison with the WACCM-Na simulations, then validated by the ground-based lidar measurements, with a focus on the geographical distribution of the layer, and the seasonal, monthly, and diurnal variations in Na density. We also examine the link between Es layers and SSLs. Sect. 4 summarizes the conclusions of this study.

80
The CMP deployed a chain of four Na resonance fluorescence lidars along 120 • E longitude at Beijing (40.2 • N, 116.2 • E) (Gong et al., 2013), Hefei (31.8 • N,117.3 • E) (Dou et al., 2009), Wuhan (30.5 • N, 114.4 • E) (Yi et al., 2013), and Haikou (19.5 • N, 109.1 • E) (Jiao et al., 2017). The transmitter of the lidar system is a frequency-stabilized dye laser pumped by a Nd:YAG laser, tuned onto the Na D2-resonant absorption line at 589.6 nm to excite resonant fluorescence from Na atoms between 75-110 km altitude. A telescope collects the backscattered photons, which are then recorded by a photomultiplier tube using an interference 85 filter centred at 589 nm. The CMP lidars can only provide night-time measurements since they do not operate in daytime. Up to January 2021, the four Na lidars have provided the long-term routine measurements for a total of 2,136 nights (19,587 hours). Figure 1 shows the monthly variations in the number of observations measured by the four lidars. Because clear weather is required for Na lidar measurements, and there is a regular presence of convective weather and thunderstorms during summer, there is a higher measurement coverage during winter. Therefore, the number of valid Na observations from lidars exhibits a 90 seasonal variation. Table 1 summarises the Na lidar data used in this study as well as the lidars' primary parameters.
Odin is a limb-scanning satellite co-funded by Sweden, Canada, France, and Finland (Murtagh et al., 2002). On February 20, 2001, the satellite was launched from Svobodny, Russia. It is in a sun-synchronous orbit at approximately 600 km, with a descending node at 06:00 local time (LT) and an ascending node at 18:00 LT. The satellite conducts limb scans from 10 to 110 km altitude. The OSIRIS spectrometer is one of the two main instruments onboard (Llewellyn et al., 2004). The instrument 95 measures radiance from the limb at wavelengths between 280 and 800 nm. The profiles of mesospheric Na number density have been retrieved from the limb-scanning measurements of the Na D dayglow at 589 nm with an altitude resolution of 2 km (Gumbel et al., 2007;Hedin and Gumbel, 2011), using an optimal estimation method (Rodgers, 2000). Figure 2 shows the timelatitude distribution of the number of the Na observations from OSIRIS during 2004-2009, using a 5 day × 5 • latitude grid as a resolution. The satellite orbits do not fully cover latitudes greater than 85 • . Furthermore, mesospheric dayglow measurements 100 cannot be made at mid-to high-latitudes in the winter hemisphere, due to the lack of daylight when the solar zenith angle is larger than 92 • . Therefore, the ground-based Na lidars, in addition to OSIRIS limb-scanning radiance measurements, provide an important measurement of local Na layers, notably in filling in data gaps during the winter. A global atmospheric model of the meteoric metals Fe, Na, K, Si, Ca, Al, Mg and Ni (e.g., Feng et al., 2013;Marsh et al., 2013;Plane et al., 2014;Langowski et al., 2015;Plane et al., 2016Plane et al., , 2018Daly et al., 2020;Plane et al., 2021) has 105 been developed, in order to advance understanding the meteor astronomy, atmospheric chemistry, and transport processes that control the different metal layers in the MLT (Feng et al., 2015;Plane et al., 2015;Wu et al., 2019). The model uses a seasonally varying meteoric injection rate of these metals at different latitudes and altitudes. In the present paper, we use a version of WACCM-Na (Marsh et al., 2013) nudged with NASA's Modern-Era Retrospective Analysis for Research and Applications (MERRA2) (Molod et al., 2015). Here we run the model with a very high vertical resolution (144 vertical levels) There is strong evidence for close coupling between the metal layers and ionospheric Es layers (Xue et al., 2017;Qiu 115 et al., 2018). Es layers are thin-layers of highly concentrated plasma composed of metallic ions and electrons, that occur between 90-130 km altitude. At mid-latitudes they are caused by the vertical convergence of ions at the null points of wind shears (Whitehead, 1960;Davis and Johnson, 2005;Chu et al., 2014;Yu et al., 2021c). Strong plasma irregularities have a considerable impact on the amplitude and phase of radio occultation (RO) signals from the global navigation satellite system (GNSS) (Yue et al., 2015(Yue et al., , 2016. Therefore, the ionospheric effects on GNSS RO signals can be used to investigate the global 120 ionosphere's morphology. The GNSS RO data from the Constellation Observing System for Meteorology, Ionosphere and Climate (COSMIC) satellites during 2006-2014, have been used to study Es layers (Yu et al., 2020). The maximum S4 index occurring between 90-130 km can be used as a proxy for the intensity of an Es layer (Yu et al., 2019b(Yu et al., , 2021bQiu et al., 2021a). In the present study we also looks into the climatology and seasonal variability of metallic ions within the Es layers, as well as the link between Es layers, the Na layer, and the presence of SSLs at low-and mid-latitudes.

125
3 Results and Discussion 3.1 Global map and seasonal variation of Na layers Figure 3 shows the time-latitude distribution of the Na column number density from the OSIRIS limb-scanning radiance measurements between 2004 and 2009, with a resolution of 5 day × 5 • latitude grid. The Na column density is integrated from 76 to 106 km altitude. Grey-shaded areas represent regions where OSIRIS limb-scanning radiance measurements were 130 not made in winter. The time-latitude distribution of the Na column number density from OSIRIS shows a distinct annual cycle of the Na column density at mid-and high-latitudes, consistent with previous ground-based and satellite measurements (Plane et al., 1999;Fan et al., 2007b;Li et al., 2018). At latitudes between 30 • and 90 • , the annual variation in Na column density becomes very significant. The maximum Na column density during winter is approximately 6×10 9 cm −2 , and the summer minimum is approximately 1×10 9 cm −2 . At low latitudes between 0 and 30 • , a small semi-annual variation in the 135 column abundance in seen with maxima of 5×10 9 cm −2 in May and October. The stellar occultation measurements made by the GOMOS instrument onboard the Envisat satellite also showed a clear semi-annual oscillation in Na vertical column density at low latitudes that merges into an annual variation above 30 • latitude (Fussen et al., 2010;Langowski et al., 2017). In the present study, the latitude dependence of the seasonal variation in the Na layer from the meridional chain of the CMP Na lidars is discussed in section 3.2. radiance data from OSIRIS provide a near-global view of Na layers. However, because the peak Na density was not observed during the dark winter months, the annual mean Na column abundances at high latitudes from OSIRIS cannot be made.
The maps in Figure 5 show the global distributions of the Na number column density from OSIRIS for four different seasons in a 5 • × 5 • grid. The Na layer clearly shows a global seasonal dependence with a pronounced minimum in the summer hemisphere. The Na column density at high latitudes reaches over 5×10 9 cm −2 in autumn and winter, and decreases 155 to 1×10 9 cm −2 in summer. High concentration above 4.5×10 9 cm −2 in all four seasons can be seen over eastern Asia and the north Pacific, the north Atlantic, and the south Pacific and south Atlantic. Note that grey-shaded areas represent areas where OSIRIS limb-scanning radiance measurements were not made during the winter at high latitudes. Whereas, the ground-based lidar measurements are a supplement to satellite data sets when the high-latitude Na layers are not solar-illuminated in winter.
The Na column density has a maximum in winter from worldwide lidar observations (Plane et al., 1999;Yuan et al., 2012;Li 160 et al., 2018) and global modeling studies of meteoric Na in the MLT (Marsh et al., 2013;Wu et al., 2021). The lower panels during each season depict the northern and southern polar views of the Na column density, and these views make the summer minimum in the high-latitude Na layers clearer. The significant summer-time depletion of high-latitude Na layers is primarily attributed to the very low temperature at the summer mesopause because of dramatic adiabatic cooling of upwelling air (Plane, 2003) and secondarily attributed to efficient removal of metallic species on noctilucent cloud particles (Plane et al., 2004;165 Raizada et al., 2007;Plane, 2012).
To compare with the OSIRIS results, the global map of the annual mean Na column number density simulated by WACCM-Na from 2004 to 2009 is shown in Figure 6. The WACCM simulated Na column density ranges from 3.0×10 9 cm −2 to 4.6×10 9 cm −2 . The largest Na column densities are seen at 60 • S high-latitude, where the column density is 4.7×10 9 cm −2 .
In the tropics, the Na column is relatively small around 3.0×10 9 cm −2 . The WACCM-Na model prediction has been shown in 170 good agreement with the Na column density measured by the mid-latitude lidar at Fort Collins (41 • N, 104 • W) and the lidar at the South Pole (Marsh et al., 2013). Figure 7 show the global distributions of the WACCM simulated Na column density for four different seasons during the period 2004-2009. The Na column density in WACCM-Na shows a similar seasonal variation to the OSIRIS observations in Figure 5. At high latitudes, the summer minimum is around 2.0×10 9 cm −2 in the WACCM-Na model, which is 175 consistent with the measurements from OSIRIS in summer. The winter maximum in WACCM-Na is 5.0×10 9 cm −2 at high latitudes. However, the ares of pronounced Na column density observed from OSIRIS shown in Figures 4-5, e.g. in eastern

The maps in
Asia and the north Atlantic, are not reproduced in the WACCM-Na simulations. In the following section, the ground-based measurements of Na layers from four lidars in the CMP are compared to satellite observations and WACCM simulations. There are over 9,000 hours and 5,000 hours observations of the Na layer at Beijing and Hefei, respectively, and more than 1,800 hours at the other two stations.
185 Figure 9a shows the comparison of annual mean Na column density between OSIRIS measurements and Na reference by Plane (2010). Good agreement is found between 20 • S and 40 • N latitudes. Whereas, the annual mean OSIRIS at higher latitudes is largely underestimated due to less measurements in winter. Therefore, the height-latitude distributions of the neutral Na layer from OSIRIS between 20 • S and 40 • N latitudes and Na lidars are shown in Figures 9b and c. In Figure 9b, the annual mean Na number density from OSIRIS shows a northern number density over 3000 cm −3 at 10 • N-40 • N latitudes and a southern density 190 over 3000 cm −3 at 15 • S latitude. Figure 9c shows the Na number density from the lidars at Beijing (40.2 • N), Hefei (31.8 • N), Wuhan (30.5 • N), and Haikou (19.5 • N), along with superposed contour lines of Na number density from OSIRIS. The peak number density of Na layers at the four stations is between 2500 cm −3 and 3000 cm −3 , in accord with the observations from OSIRIS. The low-latitude Na layer at Haikou has a relatively higher peak density of nearly 2900 cm −3 than the mid-latitude Na layers at Hefei and Beijing with the peak density of 2500-2550 cm −3 . The Na layer at Wuhan has a moderate peak density of More recently, the distribution of Mg + column density simulated by WACCM-X (Wu et al., 2021) shows a stronger con-200 vergence along the magnetic equator. The equatorial metallic ions are uplifted by V×B forcing and then drift down along the magnetic field lines to the subtropical region. The transport of metallic ions is generally consistent with the fountain effect and the fountain effect is stronger over east Asia (Wu et al., 2021). The metal layer can be influenced by the metallic ions as the   It is an interesting feature that both the neutral Na layer and Es layer intensity have a similar latitude distribution (at least below 40 • N latitude). Many previous studies have shown a strong correlation between local SSLs and Es layers (Dou et al., 2010;Kane et al., 2001;Sarkhel et al., 2012;Dou et al., 2013). The neutralization of metallic ions is the most commonly accepted mechanisms for SSL formation (Cox and Plane, 1998). Na layer and Es layers both have a prominent seasonal variation. Na layer has a maximum density in winter and the Es layers have a maximum density in summer. The geographical 215 distributions of Na atoms and metallic ions with Es layers, particularly the similar pronounced areas of high concentration should be related. Therefore, it indicates that it is necessary to incorporate the electrodynamical transport of metallic ions within Es layers into a global model of Na.  depletion of the Na layer is observed at Hefei from lidar, with a minimum of 1700±100 cm −3 in July. The maximum Na density in winter at Hefei is 3100±100 cm −3 in November from Na lidar and 5000±400 cm −3 in October from OSIRIS. The depletion of the relatively low-latitude Na density in summer is not significant at Wuhan and Haikou, and the Na layers show a semiannual variation with peaks at equinox from the Na lidars. In Figure 10f, the winter maximum from Na lidar at Wuhan is 4100±300 cm −3 in January and the secondary peaks are 3000±300 cm −3 in April and 3700±200 cm −3 in November. In 240 Figure 10h, the peaks are 4300±200 cm −3 in October and 3400±300 cm −3 in March. In Figures 10e and g, the Na number densities from OSIRIS show an annual variation with maxima of 5400±400 cm −3 and 5600±700 cm −3 at Wuhan and Haikou in October. The Na number density from OSIRIS is generally larger than that from Na lidars at the four stations.
On the right panels of Figure 10, the peak densities of the Es layers in summer are superimposed on the Na density contour from lidars, represented by the S4max at levels of 0.6, 0.8, 1.0 as the light to dark blue contour lines. The response of the Na 245 layer to the strong Es layers in summer is not very significant. Even though the density of the Na layer is lowest in summer because of the very low temperature at the mesopause, the Na layer can be influenced by the neutralisation of metallic ions with a summer maximum of Es layers. It has been found that the sporadic enhanced Na layers in the upper mesosphere and thermosphere occur more frequently in summer than around the equinoxes, which are generally associated with the presence of the Es layer (Dou et al., 2013). The sudden enhancement of neutral metal density, which appears abruptly in a thin layer (full 250 width at half maximum ∼1 km) on the topside of the normal metal layer (Clemesha, 1995), has a strong seasonal dependence with the highest occurrence rate in summer but it was rarely observed in other seasons (Qiu et al., 2016). Based on the observations from lidars and satellites, the climatology and seasonal variations of the Es layers, the sporadic metal layers (SSLs for Na), and the Na layers are analysed later. Figure 11 shows the monthly variations in Na number density in the evening from OSIRIS and Na lidars. Because the time of 255 the ascending node shifted from 18:00 LT to 18:50 LT over the years, the number of Na density data in the evening, particularly in the equatorial region, is much less than that in the morning from the OSIRIS limb-scanning radiance measurements (Hedin and Gumbel, 2011). The Na number density in the evening is smaller than that in the morning. The Na number density from OSIRIS is generally larger than that from the Na lidars. In the right-hand panels of Figure 11, the observations from the four lidars show a significant summer depletion of the Na layer. The observations from Na lidars show a shift from an annual 260 variation in Na density at Beijing and Hefei to a semiannual variation at Wuhan and Haikou, which is consistent with the seasonal variation with latitude in Figure 10. The Na number density minimum is around 1500 cm −3 in summer, and the maximum density is around 2500 cm −3 . Figure 12 shows the monthly variations in the WACCM simulated Na number density at 6:00 LT in the morning and 18:00 LT in the evening. In general, WACCM-Na is capable of reproducing the seasonal variation in the Na layer. However, WACCM 265 underestimates the seasonal variability of the Na layer as observed by lidars at Wuhan and Haikou. It may arise from the fact that the seasonal variablity of vertical transport of atmospheric constituents modeled in WACCM is underestimated since gravity waves and turbulence fluctuations are not included or the equatorial plasma fountain effect is not included in model. The vertical flux of Na atoms in the mesopause region employed in WACCM-Na is much smaller than that from lidar measurements (Marsh et al., 2013). The WACCM simulations at 6:00 LT show a peak Na number density of approximately 3000 cm −3 in 270 October and a secondary peak of approximately 2500 cm −3 in March. The WACCM simulated Na density at 18:00 LT shows a similar variation. The peak density is approximately 3000 cm −3 in October, whereas the secondary peak in March is not significant. Figure 13 shows the full diurnal variation of the WACCM simulated Na number density and the lidar-observed night-time Na number density. The Na layer is enhanced after midnight in the WACCM simulations. It is consistent with the observations 275 from Na lidars that the morning Na density around 06:00 LT is 10 %-30 % larger than the evening Na density around 18:00 LT.
After sunrise, an increase of Na density on the bottom side of the layer and removal of Na density on the topside of the layer can be found in WACCM simulations. Ion-molecule chemistry plays an important role in the Na layer above 96 km altitude (Cox and Plane, 1998). The influence of the odd oxygen (O and O 3 ) /hydrogen (OH, HO 2 ) chemistry and photochemistry of the major reservoir species (NaHCO 3 ) dominates below 96 km altitude (Plane et al., 1999(Plane et al., , 2015Yuan et al., 2019;Xia et al., 280 2020). Due to the increased solar radiation and the increased NO + and O + 2 ions, atomic Na is removed off the topside of the Na layer after sunrise and transformed to Na + . The photolysis of the main reservoir, NaHCO 3 , causes an increase in Na atoms on the bottom side of the layer after sunrise. WACCM also reproduces the diurnal variation in the Na layer driven by the diurnal tidal modulations (Liu et al., 2013;Yuan et al., 2014a). The right-hand panels of Figure 13 show that lidar-measured Na density after midnight is larger than in the evening, in agreement with WACCM. 285 Figure 14 compares the seasonal profiles of Na number density from OSIRIS, Na lidars, and WACCM-Na at Beijing (a-c), Hefei (d-f), Wuhan (g-i), and Haikou (j-l). The left panels show the vertical profiles of Na number density from OSIRIS with a 2 km altitude resolution between 76 and 106 km, within a region of ± 5 • latitude and longitude square centred on each lidar station. Due to the absence of daylight when the solar zenith angle is larger than 92 • , there are fewer observations in winter than in the other three seasons. The observations from OSIRIS show similar seasonal mean vertical profiles of Na density as 290 above the four lidar stations. The maximum Na density is 4500-4900 cm −3 in autumn, with a peak height around 91 km. The peak density is 2700-3500 cm −3 during other seasons.
The seasonal profiles of night-time Na number density from the Na lidars are shown in the middle panels. In Figures 14b and h, the Na layer has a significant seasonal dependence at Beijing and Wuhan, with the maximum Na density of approximately 3000 cm −3 in winter and a peak height of 91 km and 93 km, respectively. In Figures 14e and k, the maximum Na density 295 at Hefei and Haikou is 2800 cm −3 and 3300 cm −3 in autumn, with a peak height of 91 km. Although the vertical resolution of the profiles of Na number density from OSIRIS limb-scanning radiance measurements is lower than that of ground-based Na lidars, the OSIRIS data are nevertheless in good agreement with the measurements from the Na lidars. The observations from the OSIRIS spectrometer on the Odin satellite therefore provide a reliable measurement of the Na layer. The right-hand panels show the seasonal profiles of Na number density simulated by WACCM-Na, which exhibit similar seasonal profiles.

300
The maximum Na density is 2800-3000 cm −3 in autumn, with a peak height of 88 km. The peak density is 2300-2500 cm −3 during the other seasons.
In Figure 15, the monthly variations in the climatology of Na column number density from OSIRIS and Na lidars, and the  Figure 15a shows a strong correlation between Es layers (blue line) and SSLs (orange line) at Beijing, with a summer maximum and a winter minimum. The probability of SSLs can reach up to 100 % per day (ten SSL events out of ten nights) in June, compared to approximately 6 % per day in December and January. The intensity of the Es layer is represented by the S4max index (Yu et al., 2019b(Yu et al., , 2021bQiu et al., 2021a). S4max is 0.96 in June, more than three times the value of 0.27-0.28 in 315 December and January. Both the observations from OSIRIS and the Na lidar show an annual change in the Na column number density, with a summer minimum and a winter maximum. The SSLs are more correlated to Es layers at Beijing at a higher latitude compared the other three stations. In Figures 15b-d, the probability of SSLs shows a semi-annual variation at Hefei, Wuhan, and Hefei. The probability of SSLs reaches a peak in summer and reaches a secondary peak in February and March.
The summer maximum occurrence of SSLs is consistent with the summer maximum intensity of the Es layer. The secondary 320 peak in occurrence of SSLs in February and March is consistent with the climatology of Na layers from lidaras. Yu et al. (2021c) found a winter-to-summer dynamical process of the Es layer induced by the lower thermospheric meridional circulation. The dynamical process such as the meridional transport of metallic ions with the influence of atmospheric meridional circulations is likely related to the monthly variations in the occurrence of SSLs. The formation of low-latitude Es layers is attributed to echoes from the equatorial electrojet irregularities (Whitehead, 1989). The Es layers at low latitudes are more affected by the 325 recurrent geomagnetic activity rather than the dynamical process caused by the background wind field (Whitehead, 1970;Yu et al., 2021a). Therefore, the correlation between the probability of SSLs and the density of Es layers at low-latitude Haikou station in Figure 15d is not as significant as those at mid-latitudes in Figures 15a-c.

Summary
We present a study on the climatology of Na layers at mid-and low-latitudes, from the ground-based measurements obtained 330 from a lidar network of four Na fluorescence lidars along 120 • E longitude in the CMP, the OSIRIS limb-scanning radiance measurements detected by the Odin satellite, and WACCM-Na simulations.
The geographical distribution of the Na layer shows a global seasonal dependence from OSIRIS measurements and WACCM model. The pronounced areas of high Na concentration above 4.5×10 9 cm −2 can be seen over eastern Asia and the north Pa- providing a global climatology of Na layers. The OSIRIS limb-scanning radiance measurements were not taken during the dark winter months owing to a lack of sunshine when the solar zenith angle is larger than 92 • . The CMP Na lidars provide high temporal and vertical resolution measurements of local Na layers at mid-and low-latitudes along 120 • E longitude as a supplement to space-based observations, which can present a test of OSIRIS and WACCM.
Good agreement is found between lidar observations, satellite measurements, and WACCM simulations. At mid-latitudes 345 and high latitudes between 30 • and 90 • , the Na layers from OSIRIS show a significant annual variation, with a winter maximum and a summer minimum. At low latitudes between 0 and 30 • , a semi-annual variation is found. In accord with the measurements from OSIRIS, the observations of Na number density from the four lidars show a significant annual variation at Beijing and Hefei and a semi-annual variation at Wuhan and Haikou. The Na number density from OSIRIS is slightly larger than that from Na lidars at the four stations, particularly in autumn and early winter as a result of significant uncertainties in layers observed at Wuhan and Haikou lidar stations. This discrepancy suggests the seasonal variability of vertical transport of constituents is underestimated in WACCM because much of gravity waves is not resolved in model. The close relationship between the ionospheric Es layer, the Na layer, and SSLs indicates that it is important to consider the dynamical and electrodynamical processes of metallic ions in the lower E region of the ionosphere coupling with the Na layer into a global atmospheric 355 model of metals.
Data availability. The Na lidar data are available from the Data Centre for Meridian Space Weather Monitoring Project (https://data. meridianproject.ac.cn/data-directory/) and the National Space Science Data Center, National Science & Technology Infrastructure of China (http://www.nssdc.ac.cn). OSIRIS data are available at https://research-groups.usask.ca/osiris/data-products.php. The COSMIC satellite radio occultation data are available from the CDAAC website (https://data.cosmic.ucar.edu/gnss-ro/).