Summer polar mesopause region temperatures measured in 2001 by the ALOMAR Weber Sodium Lidar

B. P. Williams$^,$, C.Y. She$^{1}$, J.D. Vance$^{1}$, K. Arnold$^{1}$, P. Acott$^{1}$, D. A. Krueger$^{1}$, and D.C. Fritts$^{2}$

Colorado Research Associates, a division of Northwest Research Associates, Boulder, CO, USA

ABSTRACT

Temperature profiles in the mesopause region over Andoya, Norway (69N, 16E)were measured on 18 days during summer 2001, as weather permitted, by the Weber sodium resonance lidar. The daily mean temperatures measured at 87km altitude were generally 10K hotter than rocket falling sphere climatology [Luebken, 1999] compiled from measurements made over 10 years. The transition to higher fall temperatures occurred 2 weeks earlier in the lidar data than the climatology. The difference between the lidar data for 2001 and the falling sphere climatology may be due to interannual variability. The higher temperatures may have implications for PMSE formation, although the large wave perturbations present make interpretation difficult.

INTRODUCTION

• The summer polar mesopause is the coldest part of the atmosphere, with temperatures reaching below 130K, in spite of being continuously sunlight. This has led to a great deal of research both on the cause for the low temperatures and to the interesting layered structures that result, such as noctilucent clouds (Gadsden and Schröder, 1989) and polar mesosphere summer echoes (Cho and Kelley, 1993).
• The continuous daytime background has made it more difficult to measure the temperatures optically and prevented the use of inexpensive passive nightglow spectrometers.
• Most of the temperature measurements have been made by falling spheres from rockets, with many campaigns over the last 20 years Lübken (1999) (hereafter, L99) and sodium lidar in the late 1980's (Lübken and von Zahn, 1991) (hereafter LvZ91) and 2000-present (She et al., 2002a).
• This paper will present the temperature measurements made in the summer of 2001 by the Weber sodium lidar at the Arctic Lidar Observatory for Middle Atmosphere Research (ALOMAR) and compare them with previous measurements by falling sphere and sodium lidar.

SODIUM LIDAR

• The Weber sodium lidar was installed at ALOMAR in the summer of 2000. The installation and subsequent campaigns were discussed in She et al. (2002a). The Weber lidar transmitter is similar to the wind-temperature sodium lidar in Fort Collins, CO (She et al., 1992) except for a unique solid state sum frequency seed laser.
• The seed laser is locked to the sodium D$_{2a}$ line using a Doppler-free sodium cell and shifted to 2 other frequencies using an acousto-optic modulator. This 100mW seed beam is combined with a 17W pulsed YAG beam in the Pulse Dye Amplifier to obtain a 1.2W pulsed beam at the seed frequency.
• The Weber Na lidar uses as receivers the two independently steerable telescopes of the ALOMAR Rayleigh/Mie/Raman (RMR) lidar, the latter having primary mirrors of 1.8 m diameter each (von Zahn et al., 2000). For the purpose of supporting the Weber Na lidar effort, each of these two telescopes was equipped by the Institute of Atmospheric Physics (IAP) with a new focal box which allows the collection of backscattered photons into two separate fiber cables. The Na photons are received from a field-of view of 0.4 mrad full width which is inclined 0.5 mrad with respect to the optical axis of the telescope.
• During the summer of 2001, both telescopes were pointed at the zenith to make temperature measurements by the sodium lidar and noctilucent cloud measurements by the RMR lidar.
• To reduce the solar background, we used magneto-optic sodium vapor Faraday filters with a passband of about 5pm, which reduce our signal level by a factor of 4. At the end of the summer, we removd the Faraday filters during the few hours of darkness near midnight on days 232-238.
• Because of the large variations in sodium density and solar background, the temperature random error from photon noise ranged from 1-10K at 87km with a typical value of 4K in daytime. Near the peak of the sodium layer at 92km, most of the error values were in the 1-2K range. The few hours of nighttime data had errors around 1K at 87km.

RESULTS

Figure 1: Daily mean temperature profiles for June, July, and first and second halves of August. The error bars give the standard deviation in the mean of the hourly lidar profiles during each day. The thick lines are the Lübken (1999) temperatures averaged for each quarter month.

Figure 2: a: A comparison of: daily mean lidar temperatures at 87 km (X) with a polynomial fit (solid line), daily mean lidar temperature for day 227 of 2000 (square), L99 temperatures (dashed line), L99 temperatures increased 10K and shifted 12 days earlier, LvZ91 results for the first and second half of August (diamonds). b: Mesopause temperature in same format with the addition of falling sphere data from 1999 (triangles). c: Daily mean sodium density at 87km with an exponential fit. d: Daily mean mesopause altitude for lidar and falling sphere with the lidar altitude range shown by vertical lines.

• The weather at ALOMAR during the summer of 2001 was extremely poor, but we were able to obtain 103 hours of observation spread over 18 days between June 12 (day of year 163) and August 26 (day 238). The number of hours of observations in each day ranged from 2 to 14 hours.
• In Figure 1 we show the mean temperature profile for each day of lidar data divided into June, July, and the first and second half of August. These are compared to the temperature profiles from the L99 falling sphere average for each quarter month. The agreement is best in June, with the lidar temperatures moving 10K higher in July and up to 20K higher at the end of August.
• The mesopause altitude measured by the lidar moves from 89 km in June to 85 km at the end of August. The L99 mesopause altitude generally stayed around 87-88km.
• In Figure 2a, we show the time series of daily mean temperatures at 87 km altitude for the sodium lidar. The error bars indicate the standard deviation of the hourly mean temperatures on that day. Most of the hour-to-hour variability seems to correspond to wave and tide activity.
• A polynomial function fits the data well and gives a similar shape to the L99 climatology, although at a temperatures 10-20K higher. If the L99 climatology is increased 10K and shifted 12 days earlier, it agrees well with the fit to the lidar data. The daily mean lidar temperature from day 227 of 2000 was higher than L99 but lower than the polynomial fit to the 2001 lidar data.
• Four falling sphere measurements from 1999 Dropps/MIDAS campaign at the Andoya Rocket Range had even lower mesopause temperatures ranging from 116-126K on days 185-194 (shown in Fig. 2b) (Schmidlin and Schauer, 2001).
• The sodium density started increasing rapidly after day 220-230, possibly in response to the rising temperatures or changing chemical concentrations.
• The 10-20K temperature difference between the sodium lidar and the falling spheres has not been explained. With daytime measurements, one has to worry about the stability of the Faraday filters, but in late August, when the lidar-falling sphere difference was the biggest, there was no significant difference in the temperatures measured by the lidar with and without the daytime filters.

DISCUSSION AND CONCLUSIONS

• Lübken (2000) compared L99 temperatures with the rocket grenade temperatures measured from 1963 to 1972 at 66N and 71N. From 80-85km altitude, the rocket grenade temperatures were 2-3K warmer on average, but this difference is surprisingly small and may be due to long-term trends.
• Clancy et al. (1994) reported temperatures measured by the Solar Mesosphere Explorer satellite at 65N that were lower than the sodium lidar measurements of LvZ91 below the mesopause and higher above the mesopause.
• Winter temperature measurements by sodium lidar and the TOTAL ionization gauge were made at nearly the same time during the DYANA campaign in February, 1990 at the Andoya Rocket Range (69N) (Lübken et al., 1994). The sodium lidar temperatures were 7K higher than those measured by the ionization gauge.
• In the last few years, the first measurements of the temperatures at the Antarctic summer mesopause have begin to yield complementary data. Lübken et al. (1999) used falling spheres to measure temperatures at Rothera (68S, 68W) that were 2-3K (7-8K) warmer in January (February) than the northern hemisphere temperatures from L99 in July (August). At the south pole, Gardner et al. (2001) used an iron Boltzmann lidar to measure temperatures that were very similar to the MSIS90 model in late January, 2000.
• The various temperature comparisons give mixed results. The sodium lidar measurements in this paper and LvZ91 tend to be warmer than the rocket based techniques. Both of the the Antarctic measurements are warmer than the northern hemisphere falling spheres measurements. It is still unclear whether these differences arise from systematic measurement errors or different sampling volumes and times for the direct comparisons; or interannual and interhemispheric variability in the case of the climatological and Arctic/Antarctic comparisons.
• To resolve these discrepancies, we are currently analyzing simultaneous sodium lidar/falling sphere measurements from a June/July 2002 rocket campaign at the Andoya Rocket Range. Preliminary temperature measurements by the SABER instrument on the TIMED satellite were 5K colder than those measured by the Fort Collins, CO sodium lidar at 40N (She et al., 2002b), but a comparison at 69N is not available yet.