1. Uniqueness of the Colorado State LIDAR Program
2. Probing atmospheric state parameters of all atmospheric layers with ground-based LIDAR
3. The narrowband Na LIDAR system at Fort Collins
4. Innovations in measurement techniques for narrowband metal LIDAR system
• Daytime measurements in the mesopause region using a Na dispersive Faraday filter
• Simultaneous temperature and line-of-sight wind measurements using a tandem acousto-optic modulator
• Continuous-Wave Sum-Frequency Generation of 589 nm Radiation
5. Scientific contributions based on sodium LIDAR observation to date
• Eight-year Climatology of temperatures and sodium density in the mesopause region
• Observed external perturbations on temperatures in the mesopause region
6. TIMED/CEDAR collaborative science
7. Science and involvement at ALOMAR
8. Figure Descriptions
9. References

• Uniqueness of the Colorado State LIDAR Program

As laser technology improves, LIDAR (LIght Detection And Ranging) has become a useful tool for probing atmospheric properties. Much progress has resulted from the need to monitor aerosol, cloud, and air polluting chemicals. More recently, the issues of global change and the associated anthropogenic causes have led not only to the need for monitoring the long-term trends of important species, such as stratospheric ozone, H2O, CO2 and CH4, and the atmospheric state parameters (temperature, density and winds), but also by necessity to improve our basic understanding of the coupling between different atmospheric layers. These objectives turned out to be more challenging than a lay person might suspect, requiring new measurement concepts and techniques as well as more committed observations. The Colorado State LIDAR program makes contribution towards this goal by carrying out both LIDAR technology development (Applied Physics) and regular atmospheric observations (Geophysics).

 
• Probing atmospheric state parameters of all atmospheric layers with ground-based LIDAR

To appreciate the essence of atmospheric layer structure and of its monitoring, we show in Fig.1 the commonly accepted annual mean temperature profile which divides our atmosphere into layers. Solar radiation is absorbed at the XUV wavelengths to photoionize atoms and molecules above 100 km, at the UV wavelengths to photodissociate ozone at ~ 50 km and at the IR and visible wavelengths to heat the Earth's surface. This leads to high temperatures at these altitudes and lower temperatures in between, thus dividing the atmosphere into layers called, tropo-, strato-, meso- and thermo-sphere with temperature extremes between these layers termed "-pause" of the layer below it.

Since temperature plays such a fundamental role in the understanding of the atmosphere, it is of interest to devise a ground based observation system capable of measuring temperatures from ground to the thermosphere. This then was the goal of our program. At the beginning of 1980's when we started our LIDAR work, Rayleigh scattering LIDAR has been used to measure temperature profiles between 30 and 60 km in altitude (to 80 or 90 km using a high power laser and a large telescope). In this case, the air density profile is measured from Rayleigh scattering signals and the temperature profile is calculated by assuming both hydrostatic and local thermodynamic equilibrium [Hauchecorn and Chanin, 1980]. This method fails in both lower and higher altitudes due respectively to aerosol interference and to the lack of signal (inverse z2 dependence). We therefore conceived of new LIDAR techniques to address these two regions. Unlike the existing methods at the time, our new LIDAR techniques require the use of narrowband transmitter and/or narrowband receiver for measurements.

In the lower stratosphere and troposphere, we depend upon the process of Cabannes scattering by atmospheric molecules with a typical bandwidth of roughly 2-3 GHz for atmospheric temperature measurements. The narrow spectrum of Cabannes scattering is the central peak of the Rayleigh scattering spectrum which also contains sidebands called rotational Raman spectra [She, 2000]. This central peak results from density fluctuations and its cross-section is temperature and pressure-dependent, thus enabling accurate vertical profiling of these atmospheric state parameters. Unfortunately, in the lower atmosphere, laser light also scatters aerosols, called Mie scattering which has a bandwidth of less than a few tens of MHz. Since spectrally both aerosol and molecular scattering are centered at the incident laser frequency, a method to effectively disentangle them must be found, if atmospheric state parameters and aerosol properties are to be measured independently and simultaneously. We proposed a Cabannes-Mie LIDAR, using a narrowband frequency-stabilized laser and a narrowband resonance vapor (notch) filter (RVF), to separate Mie (aerosol) scattering from Cabannes (molecular) scattering, thus measuring both aerosol and temperautre profiles. Although temperature profiling in these altitudes are being measured routinely twice a day in few stations world-wide by conventional methods (balloon and radiosonde), the LIDAR measurements may be automated and be carried out more frequently with higher spatial resolution. Such temporal and spatial resolution are needed for the understanding of atmospheric motions (dynamics). With on and off funding, it took us ~ 12 years to complete the development of a system capable of measuring temperatures from 1 km to 15 km [Hair et al., 2000] using a relatively low power laser and modest telescope (dia. 14 inches). Though we no longer investigate the narrowband Rayleigh-Mie Lidar, Dr. John Hair, now at NASA Langley Research Center is continuing to pursue such system fro space application.

For altitudes above 80 km, we are in the atmospheric region with neutral alkali atoms, where not much data existed and for this reason, it was referred to as the Ignoroshere. Since resonance scattering from Na atoms is 14 orders more efficient than Rayleigh or Cabannes scattering, the process of laser-induced fluorescence (LIF) can be used to produce a strong LIDAR signal from the mesopause region even with a modest LIDAR system such as ours, deployed at the Christman field of Colorado State University. Since the narrow Na hyperfine broadend ground state spectrum is temperature and wind dependent, we use a narrowband  system to indirectly probe the details of the LIF spectrum to assess atmospheric temperatures and line-of-sight winds.

Although our initial goal was to develop a ground-based LIDAR system to probe the state of the atmosphere from ground to ~110 km, the overwhelming success of the narrowband Na LIDAR program has resulted in shelfing the continued development on the narrowband Cabannes-Mie LIDAR and in shifting our concentration to the technical development of and to upper atmospheric science study using this still developing narrowband Na LIDAR system.

• The narrowband Na LIDAR system at Fort Collins

To understand the measurement technique based on LIF, we need to know something about the atom in question, in this case sodium. A simplified Na energy diagram is shown in Fig.2 (a) with temperature-dependent LIF spectra shown in Fig.2(b). Using a laboratory Na cell, Doppler-free fluorescence spectrum as shown in Fig.2(c), which may be measured with a tunable single-frequency laser, shows sharp features at the D2a peak and cross-over frequencies respectively at na and nc. Since LIF at these two frequencies are very sensitive to atmospheric temperature changes as can be seen in Fig.2(b), the ratio of LIDAR returns at nc to na may be processed to measure atmospheric temperatures in the mesopause region (80 - 110 km) where naturally occurred Na atoms reside.

Initially, the narrowband Na temperature LIDAR system at Colorado State uses a Celestron 14 telescope as receiver. The transmitter pulses at a repetition rate of 20 Hz, a measured bandwidth about 120 MHz and an output average power of 0.6 W, were generated from a pulsed dye amplifier (PDA), pumped by a double Yag laser and seeded by a c.w. dye laser system, which is in turn pumped by an argon-ion laser (later in 1997 replaced by a c.w. Yag laser). To measure temperatures, the c.w. single-frequency dye laser is tuned alternatively to two frequencies (at na = - 651.4 MHz and nc = 187.8 MHz relative to the  589.158 nm Na D2 transition). Despite a very modest power-aperture (PA) product of 0.06 Wm2, this LIDAR system has been the work horse for many years, producing temperature and Na density measurements in the mesopause region resulting in new insights and significant geophysical implications. Typically, the resolution of our narrowband Na LIDAR is processed to give 4.7 km resolution with typical measurement uncertainties for the hourly mean temperatures of ~0.6 K near the sodium peak.

To measure temperature and line-of-sight wind simultaneously, three intensities, forming two intensity ratios at frequencies selected to acquire temperature and wind sensitivities must be measured. The required frequency agiling is done by locking the c.w. dye laser at the D2a peak, na, and use a tandem acousto-optic modulator (AOM) to shift 630 MHz sequentially to higher and lower frequency sides, giving two additional frequencies as n+ = na + 630 MHz and n- = na - 630 MHz as shown in Fig.2(c). Using the same laser power for the transmitter and telescope aperture, the measurement uncertainties for temperature and wind at the Na peak are, respectively, 0.6 K and 1.5 m/s. With the AOM permanently installed in the LIDAR transmitter, the Fort Collins LIDAR facility may be operated either in a 2-frequency or a 3-frequency mode, measuring, respectively, temperature and Na density, and temperature, line-of-sight wind and Na density. On campaign based, Faraday filter (to be discussed later) is being used to permit daytime observation.

• Innovations in measurement techniques for narrowband metal LIDAR system

With the continued support from the NSF CEDAR program, three upgrades have been pursued, developing (1). a narrowband sodium Faraday vapor filter in the receiver for daytime measurement capabilities, (2). a tandem acousto-optic (AOM) modulator in the transmitter for simultaneous temperature and line-of-sight wind measurements, and (3). a physically small c.w. solid-state laser source at 589 nm, by the technique of sum-frequency generation, to replace the bulky argon-ion and c.w. dye laser system. The significance of these developments are discussed further.

Daytime measurements in the mesopause region using a Na dispersive Faraday filter

In order to probe the mesopause region in the daytime, a ultra-narrowband filter with a bandwidth between 1 and 100 GHz is needed to reject sky background and to pass the desired signal at the same time. That a circular birefringent, dichroic medium between crossed polarizers would work was demonstrated before the laser was invented. Later, a Faraday filter with a magnetic field of 1500 G was constructed by Agnelli et al. [1975] for the observation of solar sodium D lines. Their filter gave a fullwidth of ~ 15 GHz and a maximum transmission of 25%. During the late 80's, un-related to sodium D lines, the interest in Faraday filters have been revived for the detection of narrowband laser radiation for defense purposes.

The operational principle of an ultra-narrow dispersive Faraday filter has been discussed  [Yeh, 1982] in the literature. Our group proposed and tested a Faraday filter for the Na system [Chen et al., 1993] with a 0.76 cm long cell in an axial magnetic field of 1750 G and a 189oC oven between crossed polarizers. This resulted in a band-pass filter of 86% peak transmission and 1.9 GHz FWHM. At Colorado State, we have since constructed a 1 inch long sodium cell in a magnetic field of 1800 G in an oven operated at a lower temperature of 168o C. The cell windows in a cooler oven have less chance of being attacked by sodium atoms in the cell, thereby prolonging the life of the Faraday filter, an absolutely necessary requirement for routine LIDAR operation. We have incorporated this Faraday filter into our LIDAR receiver for day time operation. Such a receiver, as shown in Fig.3, is termed Faraday receiver (as opposed to the Regular receiver without a Faraday filter).

As discussed in a publication [Chen et al., 1996], in the Faraday receiver, the collected light from the telescope passing through an interference filter is fiber-coupled into a light-tight box containing the sodium dispersive Faraday filter and a photomultiplier (PMT). The effectiveness of Faraday filter in rejecting sky background may be appreciated by the high quality data obtained as shown in Fig.4. These are hourly mean photocount profiles taken at Fig.4(a) and Fig.4(b) near local midnight, Feb. 8, 1997, with Regular and Faraday receivers respectively and simultaneously, and Fig.4(c) around the following high noon with the Faraday receiver. Compared to the Regular receiver, the detected signal in the Faraday receiver is expected to be reduced by a factor of ~10 as the measured profiles show. The detected photocounts between 120 and 155 km are from background sky light. The mean background in the Faraday receiver at midnight is essentially nil as shown in (b) and even near high noon, it is seen to be lower than that detected in the Regular receiver at night!

Based on the first principle and measured Faraday filter function, Na density and temperature for the Faraday receiver may be deduced from the 2-frequency LIDAR return. Figures 5a and 5b show an example of hourly mean sodium density and temperature profiles measured simultaneously with Regular and Faraday receivers at night on 6.5 hr. UT, Feb. 8, 1997. The fact that these temperature profiles measured at night are in agreement within the statistical error bars provides confidence on our daytime temperature measurements when Regular receiver no longer can operate. The hourly mean profile taken near the following high noon, derived from the photocount files, is shown in Fig.5(c). Several campaigns with more than 24 hours continuous observation have been completed; these data are being used for tidal analysis [Krueger et al., 1997] and study of seasonal variations [Chen et al., 2000].

Simultaneous temperature and line-of-sight wind measurements using a tandem acousto-optic modulator

Studies of the middle atmospheric dynamics are incomplete without detailed knowledge of its wind structure. Since the absorption frequency (mean wavelength lo) of an atmospheric Na atom with radial velocity VR is shifted by VR/lo, radial winds can also be measured by a narrowband LIDAR provided that Doppler shift, in addition to Doppler broadening of atmospheric Na atoms, is monitored. In order to measure both temperature and radial wind, an additional fluorescence ratio excited at frequencies sensitive to Doppler-shift (wind speed) must also be derivable from the LIDAR return. After initial attempts [Bills et al., 1993; She and Yu, 1994], the method that emerges uses a tandem acousto-optic modulator (AOM) we developed to tune among three frequencies selected. In operation, the laser is locked at the Na D2a peak frequency (na). The tandem AOM shifts the output sequentially to two additional frequencies (n± = na ± 630 MHz) necessary for the wind/temperature measurement. As depicted in Fig.6, the tandem-AOM consists of, in order, a polarizing beam splitter, a periodic series of lenses and AO crystals, a quarter-wave plate, and a back mirror to achieve double-pass shifting by either frequency-shifting crystal into the same exit beam path.  A mechanical chopper-wheel before the back mirror selects which beam order (shifted or unshifted) is allowed to traverse the unit, provides the primary synchronization, microwave power to the crystals at the correct time, and triggering the pulsed transmitter and receiver systems.  The sequence of operation is also shown in Fig.6. Top, with neither crystal powered, c.w. light from the 589.18 nm ring dye laser passes the system unshifted.  Middle, with the second crystal powered, the c.w. light exits the system in the same beampath with a frequency upshifted by 630 MHz.  Bottom, with the first crystal powered, the frequency is downshifted by 630 MHz.  Frequency shifting by the crystals occurs when a crystal is powered, in synchronization with the firing rate of our Nd:YAG laser. The shifted and unshifted light is amplified by a pulsed dye amplifier (PDA).  The PDA output consists of the narrowband high peak power, sequentially tune to the returns at the three frequencies, na, n+ = na + 630 MHz and n- = na - 630 MHz on a pulse-to-pulse basis to avoid contamination due to Na density variations. The LIDAR signal returns may be manipulated into ratios sensitive to both temperature and radial wind, Rw and RT, from which temperature and wind may be simultaneously determined using calibration curves derived from the theoretical laser induced fluorescence of Na atoms [She and Yu, 1995]. Using Na Faraday filter in receiver, simultaneous temperature and wind, thus in principle heat and momentum fluxes, can be measured in daytime as well.

With this tandem AO system we opted to point the LIDAR at the zenith for initial measurements at Colorado State;  the higher accuracy required for vertical wind measurements provides a more stringent test for this LIDAR system. The system has been in routine operation. Typical  nighttime observation yields several simultaneously measured hourly mean temperature and vertical wind profiles.  Eight hours of data (4 - 11 hr, UT) were taken in March  18, 1997 and three sets of the profiles are shown in Fig. 5. It is clear that the hourly temperature display expected range of temperatures and variations with error bars around ± 0.5 K. The corresponding wind error is about ± 1.5 m/s. To further reduce the wind uncertainty so that the vertical wind (which is believed to have smaller geophysical variability) can be properly determined, a much larger receiving telescope, such as the one at the ALOMAR observatory will be needed. What is more disturbing is perhaps the fact that the hourly mean vertical winds (including those hourly profiles not shown in Fig.5(b) ) near the sodium peak show a systematic downward bias of roughly 6 - 9 m/s, a value considerably larger than expected, indicative of an instrumental problem.

We note that the LIDAR transmitter beam is the output of a pulsed dye amplifier (PDA) which is injection seeded by a single-mode c.w. dye laser tuned to the Na D2 resonance line.  Up to this point, we have assumed the centroid frequency of the pulsed output is identical to the frequency of the c.w. dye laser which serves as the frequency marker [She et al., 1992] via Doppler-free fluorescence spectroscopy [She and Yu, 1995]. Calculations show [White et al., 1996] that if the centroid frequency of the pulsed output is blue shifted from that of the c.w. laser by 13 MHz, a bias in the vertical wind velocity of -8 m/s, similar to those shown in Fig.5, results. The same frequency shift will give a temperature bias of less than 0.5 K. The simultaneous temperature and line-of-sight measurement technique has therefore been implemented for our routine LIDAR operation, yielding accurate temperature measurements for our continued long-term study of the thermal structure of the mesopause region [She et al., 1998]. To verify the existence of the blue shift, we have compared the difference in absorption in an iodine cell between the c.w. light and its pulsed output. Laboratory experiments have shown this to be the case and we are now in the process of developing a device based on an iodine cell and employing the edge technique [Korb et al., 1992] to monitor this frequency shift in real time. In this manner, even if the conditions (pump power and dye quality) of the PDA are changed during data acquisition, the corresponding frequency shift can be determined and its effect properly assessed and corrected for. When this is implemented, the LIDAR return may be adjusted to provide accurate temperature and line-of-sight wind measurements to within 0.5 K and 1 m/s accuracy.  With both vertical winds and temperatures precisely measured, the profile of heat flux [Tao et al., 1996], for example, can be experimentally determined.

Continuous-Wave Sum-Frequency Generation of 589 nm Radiation

Nature provides the coincidence that the sum frequency generation with two Nd:YAG lasers, one near 1.06 mm the other near 1.32 mm, produces the light at the sodium D2 transition (i.e. 589 nm).  It is even more remarkable that both lasers are commercially available and can be operated near the maximum of their respective tuning curves while generating light at 589 nm.  This coincidence has been used by researchers at M.I.T. [Jeys et al., 1989] to generate high power, pulsed sodium resonance radiation with two such pulsed Nd:YAG lasers for adaptive optics applications.  However, their system does not have the bandwidth and frequency control necessary for temperature and wind measurements.  For these measurements it is necessary to have a very narrowband light source ( ~ 100 MHz) and to be able to control its center frequency to an accuracy of a few MHz, ideally locking it to one of the sub-Doppler features of the sodium spectrum.  As discussed earlier such a light source should incorporate the cw generation of narrow-band sodium-resonance radiation, which can be amplified in a pulsed dye amplifier (PDA). Adding to the need for diagnostics, a c.w. power of 200 - 300 mW is sufficient for LIDAR operation.

We had indeed proposed and received funding to research the use of two monolithic c.w. Nd:YAG lasers, one operating at 1.06 mm and the other one at 1.32 mm to generate sum-frequency at 589 nm in a LiNbO3 crystal. In two years, Moosmuller (CSU Ph.D.) and Vance [1997] of University of Nevada generated ~2 mW light at 589 nm from a simple LiNbO3 crystal using 700 mW at 1.06 mm and 300 mW at 1.32 mm. More recently, with a doubly resonant congruent lithium-niobate resonator in an arrangement shown in Fig.7, Vance, She and Moosmuller [1998] had demonstrated a maximum of 400 mW of power at 589 nm using the same monolithic input lasers. The system can be operated at 300 mW level with the output frequency locked to the Na D2a peak frequency. Since the bandwidth of the c.w. output so generated from the monolithic lasers is much narrower than that of the c.w. dye laser, ~0.1 MHz vs. 1 MHz, the new c.w. light source is expected to be more stable and the resulting LIDAR more robust.

• Scientific contributions based on sodium LIDAR observation to date

Measurements of temperature profiles between 80 and 110 km have long been a challenge. The scarcity in observational data has hindered our understanding of this important atmospheric region which couples the lower atmosphere to the thermosphere. The introduction of ground-based narrowband resonance metal LIDAR [Fricke and von Zahn, 1985], with potential for measuring a temperature profile in a few minutes, has begun to drastically change the situation. A variant of this type of instrument, termed two-frequency narrowband sodium LIDAR, was developed by our group and the first LIDAR temperature profile in a midlatitude mesopause region was obtained at Fort Collins, CO in August 1989, in collaboration with the Illinois group [She et al., 1990].

Using the two-frequency narrowband Na temperature LIDAR, considerable new geophysical observational results have been reported by the Illinois [Bills and Gardner, 1993] and the Colorado State groups. Contrary to COSPAR International Reference Atmosphere (CIRA 1986) predictions [Fleming et al., 1991], two prevailing temperature minima were observed [She et al., 1993] in the nighttime temperature profiles, suggesting the significance of either wave or exothermic chemical heating mechanisms in the mesopause region. Gravity wave parameters including the Brunt-Vaisala frequency can and have been determined directly from the measured temperature profiles [She et al., 1991]. Seasonal temperature variations and the nighttime temperature climatology of a midlatitude mesopause region have been published [Senft et al., 1994; Yu and She, 1995]. Benefiting from comparable latitudes, the Na temperature LIDAR data taken at Fort Collins (41oN) covering 83 - 105 km and Rayleigh LIDAR data taken at Southern France (44oN) covering 30 - 87 km have been combined to deduce the vertical structure of the midlatitude temperatures from stratosphere to mesopause (30 - 105 km) [She et al., 1995 ]. Along with a shipborne LIDAR observation between 71oS and 54oN [von Zahn et al., 1996] using a narrowband potassium LIDAR, and observations at 41oN and 54oN during the 96-97 period, the observed two-level mesopause thermal structure [She and von Zahn, 1998] has shown to have global validity. Our Faraday filter allowed us to conduct daytime observations and a campaign over a two year span was conducted from which 18 sets of complete diurnal temperature measurement (24 hours of longer) well distributed throughout the year were obtained. The nightly means are different from the true diurnal means [States and Gardner, 2000], and our recent 18 sets of continuous diurnal observation have revealed that depending on season, approximately the nighttime seasonal mean is colder (warmer) than the diurnal annual mean by less than 5 K above (below) a nodal point in the profile where the three means (nighttime, diurnal and daytime) are nearly the same. This is shown as shown in Fig. 8 below. The annual (not shown) nighttime mean is cooler than diurnal mean by no more than 2 K below, and warmer by no more than 3 K above the annual mean nodal point, at 88 km. These differences between nighttime mean and diurnal mean do not alter the general thermal structure of the mesopause region. In fact, the concept of two-level thermal structure appears to be even more robust when diurnal means are interrogated [Chen et al., 2000].

Eight-year Climatology of temperatures and sodium density in the mesopause region

Long data sets now exist from our routine nighttime temperature measurements in the mesopause region. Along with 4 and 5 nights of initial observations, respectively, in springs of 1990 and 1991, quality temperature measurements, i.e., on average four to five nights a month with 4 hours or more observation each night, were made over Fort Collins, CO (41oN, 105oW) starting May 29, 1991. A mean temperature profile was computed each night from vertically photocount profiles smoothed by a running Hanning window, FWHM of 3.7 km, with a typical measurement precision of ~0.6 K and ~5 K near the peak (92 km) and edges (81 km and 107 km) of the Na layer, respectively. By March 30, 1999, a total of 417 nightly mean temperature profiles were obtained and carefully scrutinized, forming a data base covering a nine-year span for the study of climatology as well as climatic change. Regular nighttime observation continues.

Using this data set of nine-year span, monthly mean profiles (Hanning window with full width of 61 day) are then made for each day of the year, from which eight-year composite altitude/month contours of sodium density and temperature were made recently; they are shown in Fig.9(a) and Fig.9(b). We have made similar contour plots for individual years. The general structure, such as two-level mesopause in the temperature structure and higher sodium density in the winter months, is similar for each year. However, the detailed contours can differ considerably from one year to the next. In addition to the geophysical variability that exists, part of the difference from one year to the next can be the result of data gaps due to missing nights and/or shorter local time coverage. The eight-year composite with 417 nights of quality data should give reliable altitude/month climatology for both sodium density and temperature in the mesopause region. A paper on these climatology has been accepted for publication [She et al., 2000]. Earlier, the nocturnal hourly mean temperatures covering a 7-year span have been used to study semidiurnal tides and associated variations [Williams et al., 1998]

Observed external perturbations on temperatures in the mesopause region

More unique and we are extremely excited about the fact that our long data set has revealed, and continues to reveal unexpected effects of external perturbations on the temperatures in the mesopause region. This clearly demonstrates the existence of coupling between different layers of the atmosphere. In addition to robust climatology, the long data set can also be used to evaluate long-term climatic change. Of considerable interest is the anthropogenic effect which, according to model output, warms the troposphere and cools the middle atmosphere and thermosphere [Roble and Dickerson, 1989; Rind et al., 1990 ]. The observed cooling of ~ 1 K/Y is, however, too large to be so attributed, since it takes nearly half to a century to double atmospheric CO2 resulting from anthropogenic sources [Brasseur and Hickman, 1988].

Based on an earlier data set of 300 nights up to March, 1997, we have reported the observation of episodic warming, attributable to Mount Pinatubo eruption [She et al. 1998], and hints of solar variability affecting temperatures in the mesopause region. The time series of residual temperatures (seasonal variations removed) at 86 and 100 km and best fit to a model containing a linear trend and an episodic warming is shown in Fig.10(a), Fig.10(b), Fig.10(c), and Fig.10(d). The altitude dependence of the peak temperature increase, DTmax, and of the time this occurs, tmax, have now been published [Krueger and She, 1999]. In the same paper, we have shown clearly that the episodic warming rides on a background cooling for altitudes between 84 and 102 km. Since most of our observation in that data set took place during the falling phase of solar flux, we suggested that solar variability may be responsible for most of the observed background cooling in the mesopause region. We stated with excitement and anxiety that as the next solar maximum (cycle 23) approaches, the validity of this conjecture may be tested as our observation continues into the year 2000 and beyond. During the period of the rising solar flux, one should observe background warming instead. If this turns out to be true, the anthropogenic effect may be then deduced, in principle, by comparing the background cooling and warming rates.

Using the new data set with 417 nights of observation, we should be able to make statements that will shed light in this direction. The major features in climatic change during this nine-year period may be surmised, without much analysis, from the measured time series of nightly mean temperatures at 86 and 100 km, shown in Figs.11(a) and Fig.11(b), respectively. Because seasonal perturbations dominate the main temperature variations at 86 km, they must be removed before climate changes in periods of many years may be studied. However, seasonal variation at 100 km is minimal and we should be able to observe the temperature changes in time scales of several years qualitatively in Fig.11(b). Within the considerable geophysical variability, a simplistic description of this time series could be a warming episode, peaking in 1993, riding on a cooling background as can be seen for data before solar minimum at 6.49 Y. Between the solar minimum (6.49 Y) and the end of the new data set (9.25 Y), there exists no obvious episodic event. Not only the apparent cooling background observed before the solar minimum has been stopped, in fact, it appears to have turned into a warming background.

It is possible to fit the time series of seasonal variation removed temperatures, T(t), to the sum of episodic warming, solar and nonsolar responses as:

A signal indicating temperature change in response to solar flux variability is clearly there. Also, unexpected negative correlation exists above 103 km and below 83 km. Such an altitude dependence on the phase of solar response indicates dynamical effects at work.  We note with interest that these altitudes where the phase reversal in temperature response to solar variability occur are consistent with French midlatitude incoherent scatter radar data between 100 and 140 km and Rayleigh LIDAR data below 80 km, see Figs. 5a and 5b of Chanin et al. [1989]. Their Fig. 5(a) is reproduced here in the right of Fig.12, where they presented the winter and QBO-West data which has stronger solar response. Taking the data of incoherent scatter radar, sodium LIDAR and Rayleigh LIDAR together, one obtains an alternating negative and positive temperature response to solar activity as a function of altitude between 30 and 140 km, providing a strong evidence of dynamic coupling existing from troposphere to thermosphere, affecting solar activity induced signatures. Also shown in Fig.12(a) is the simulation of Huang and Brasseur [1993], giving simulated response in good agreement with observation, ~ 10 K change between solar maximum and solar minimum. The simulation failed to produce phase reversal, presumably due to their inadequate treatment of dynamical feedback effects.

The nonsolar response is within 1 K/Y as the results of other observation have indicated. The trend we deduced however showed both cooling and warming, respectively, above and below 93 km. It is interesting to note that midlatitude Rayleigh LIDAR showed a strong cooling trend between 60 and 70 km, but they change signs above 77 km, see Fig. 4a of Keckhut et al. [1995], again consistent with our results covering 83 to 105 km. Since dynamical feedback plays an important role in climatology, and in responses to episodic and solar activity, it could very well affect local temperature trends as well. In the midst of all these interesting results, we are reminded that due to relatively large uncertainty the trend observed is only of marginal statistical significance. We also point out that our result is from one location, and it should not be taken as the global mean.

These intriguing effects will be become more transparent as we complete one solar cycle observation. The unique data set we obtained at this point suggests that continued observation to cover multiple solar cyscles at Fort Collins would be extremely valuable.

• TIMED/CEDAR collaborative science

After the overwhelming success of the UARS (Upper Atmospheric Research Satellite) for stratospheric studies, NASA has decided to launch a new satellite, called TIMED (Thermosphere, Ionosphere, Mesosphere and Electrodynamics), for upper atmospheric studies with core altitudes between 60 and 180 km. With four instruments on board, GUVI, TIDI, SABER and LEE, there are still many important research questions left un-covered. Thus, correlative ground-based measurements are essential to extract the maximum science benefit from the space-borne observation. Since our long-term probing of the mesopause region has demonstrated the unique capability for measuring temperature and line-of-sight wind in the mesopause region, in both daytime and nightime, we proposed to upgrade our narrowband Na LIDAR to a two-beam system, pointing East or West and North or South, so that measurements of temperature, zonal and meridional winds can be made simultaneously, with both daytime and nighttime capabilities. This upgrade, funded by SNF CEDAR program has been completed.

We believe that our continued observation, starting 1991, will be able to provide a unique basis from which to evaluate the data from the proposed synergistic observations with TIMED instrumentation to determine the global thermal structure of the mesopause region. Thus, we proposed, for a two-year science, regular observational campaigns for validation as well as synergistic observation measuring temperature, zonal and meridional winds, with the space-borne measurements, SABER and TIDI. This will allow us to determine both tidal components and mean winds and temperatures unambiguously, contributing to one of the main objectives of the TIMED’s large-scale waves and dynamics subgroup. Observations for the proposed research will typically consist of 3 full diurnal cycles per month. Along with two more nights per month from an ongoing NSF supported project, high quality LIDAR data collected should be enough for establishing the climatology of temperature and winds in the mesopause region. Because of its implication on energetic balance and transport associated with the diabatic circulation resulting from gravity waves from lower altitudes, one of our observational emphases will be the zonally averaged meridional winds.

In addition to the two-beam narrowband Na LIDAR, we proposed the formation of a cluster of optical instruments located in the Fort Collins area. These passive instruments, with horizontal coverage and much higher temporal resolution, consist of an all-sky imager (Whitworth College) for imaging horizontal wave structure, an OH Temperature Mapper (USU) for imaging horizontal temperature structure, and a MODI (magneto-optic Doppler imager; CSU) for imaging horizontal structure of sodium nightglow intensity plus mean winds in the mesopause region. These “unattended?instruments will be able to observe a typical 10 nights per month. With the background winds provided by the LIDAR at Fort Collins and MF radar at Platteville, intrinsic phase velocities of waves observed by the imagers can be determined. The cluster of instruments will measure fluxes of energy and momentum of upward propagating gravity waves, as derived from perturbations in OH emission intensities, OH temperatures, and Na winds, directly serving the key scientific objective of TIMED’s small-scale waves subgroup.  The cluster in the Fort Collins, CO will be connected with other clusters in the Starfire range, NM and Bear Lake, UT, via overlapping all-sky imagers to access and probe the dynamical structures over the entire Rocky Mountain region.

• Science and involvement at ALOMAR

Fort Collins is a midlatitude site. In order to obtain information on global atmospheric structure, one needs to measure the mesopause region over equatorial as well as polar latitudes. We are collaborating with Arecibo observatory (18N) and assisting Dr. Friedman’s (CSU Ph.D.) effort on a narrowband potasium LIDAR. Our long collaboration with Prof. Von Zahn, a co-founder of ALOMAR (Arctic LIDAR Observatory for Middle Atmosphere Research) observatory has led to our deployment of a state-of-the-art narrowaband Na LIDAR transmitter based on our innovation of AOM, Faraday filter and sum-frequency generation of 589 nm laser-like light. We are deploying such a transmitter in the summer of 2000. Faraday filter is absolutely necessary for the observations in Arctic summer under 24 hour sunlit condition. To measure temperature and wind at a polar region is one of the reason we go to ALOMAR, Norway (69N, 12E). The main reason however is their twin 1.8 m dia telescope there, which will greatly increase signal-to-noise and enhance our measurement capability. In addition, the collocated instrumentation, a Rayleigh LIDAR, an ozone LIDAR, a MST radar and a MF rada plus a number of air glow instruments, as well as rocket launching capability would complement our measurements with Na LIDAR to investigate the dynamics of gravity waves which govern and provide quantitative understanding on the behavior of cooler summer and warmer winter in the mesopause region. To appreciate the uniqueness of ALOMAR observatory and the important role our Na LIDAR will play, we quote the words of our main collaborator and adviser of our ALOMAR project, a reknowned gravity wave specialist, Dave Fritt of Colorado Research Associates below.

The Faraday filter and tandem AO modulator (items 1 and 2) have been integrated into the LIDAR system since 1996 and quality data for daytime temperature and for nighttime simultaneous temperature/wind measurements have already been made. In collaboration with Moosmuller at Nevada University, the solid-state laser source at 589 nm have been developed and tested to have sufficient power. Its integration into the Colorado State LIDAR is planned for this coming summer. The proposed sodium LIDAR system to be deployed at different stages at the ALOMAR observatory will contain these exciting upgrades. The narrowband Na LIDAR with these upgrades implemented would be capable of simultaneous measurements of temperature and line-of-sight wind, day and night, weather permitting. The potential scientific pay-off is expected to be great.

Tidal analysis based on the same 18 data sets covering complete diurnal cycle have been made and diurnal and semidiurnal amplitudes and phases for the 4 seasons determined. Even though the available data sets may be too small to average out long period gravity waves and one can not separate non-migrating tides from migrating tides with data from a single station, the observed seasonal mean tidal phases appear to be in general agreement with GSWM (Global Scale Wave Model) predictions for migrating tides [Hagan et al., 1995]. Depending on season, the height profiles differ, but the observed tidal amplitudes, 4 ?12 K are in the same range of predicted values. An article on our tidal results in comparison to both TIME-GCM [Roble and Ridley, 1994] and GSWM simulations is being prepared for publication.

The unique and comprehensive suite of instruments in ALOMAR will enable a number of quantitative and compelling studies of dynamics extending from the major wave source regions at lower altitudes into the MLT (mesosphere and lower thermosphere). Thus, for the first time, studies that address the coupling of wave sources and variability at lower altitudes to responses and variability at MLT altitudes will be possible. Also enabled will be quantitative studies of the wave-wave and wave-mean flow interactions, coupling, and dynamics within the MLT that contribute additional variability at these altitudes and themselves influence the structure and dynamics at even greater altitudes.

 No other observational facility on the globe is now capable, or envisioned to have the capability in the near future, of measuring winds, temperatures, heat and momentum fluxes, and constituents nearly continuously throughout the atmospheric column from the troposphere into the lower thermosphere. Indeed, these combined measurements will yield unprecedented opportunities for quantitative and comprehensive studies of many aspects of lower atmospheric and MLT coupling and dynamics important to the scientific community as a whole.

Perhaps the most unique capabilities of the enhanced ALOMAR instrumentation will be the potential to address various aspects of gravity wave propagation, filtering, wave-wave and wave-mean flow interactions, spectral evolution, and dissipation processes that can only be done in part or not at all at other present and planned measurement facilities.

Figure Descriptions

Fig. 1. Annual mean temperatures of atmospheric layers.

Fig.2(a), Fig.2(b), and Fig.2(c). Sodium D2 transition and narrowband fluorescence LIDAR: (a). simplified energy structure of a sodium atom, (b). temperature-dependent Doppler broadened lineshapes, and (c). Doppler-free fluorescence features at ~320 K and LIDAR transmitter frequencies, and measurement intensity ratios.

Fig.3. Schematic of the Faraday receiver channel with a fiber coupled sodium-vapor dispersive Faraday filter, capable of rejecting considerable daytime sky background.

Fig.4(a), Fig.4(b), and Fig.4(c).   Hourly mean photocount files taken at Fort Collins, CO (41oN, 105oW) on Feb. 8, 1997 around local midnight, (a) and (b), with Regular and Faraday receivers, respectively, and around high noon, (c), with the Faraday receiver.

Fig.5(a), Fig.5(b), and Fig.5(c). Hourly mean Na density (a) and temperature (b) profiles taken simultaneously near local midnight, Feb. 8, 1997, with Regular and Faraday receivers simultaneously, and the hourly mean temperature profile (c) taken the following high noon with the Faraday receiver.

Fig.6. Diagram of the tandem acousto-optic modulator unit. Top, with neither crystal powered, c.w. light at 589.18 nm passes the system un-shifted. Middle, with the second crystal powered, the c.w. light exits the system in the same beam-path with a frequency up-shifted by 630 MHz. Bottom, with the first crystal powered, the frequency is downshifted by 630 MHz.

Fig.7. A doubly resonant congruent lithium-niobate (LiNbO3) resonator for sum frequency generation, showing paths of input beams at 1.06 mm and 1.32 mm as well as the generated beam at 589 nm.

Fig.8.  Seasonal nighttime, daytime and diurnal mean temperature profiles based on 18 contiuous diurnal observations distributed throughtout the year.

Fig.9(a) and Fig.9(b). Eight-year climatology of 3.7 km and 1 month smoothed nocturnal temperature (upper) and Na density (lower).

Fig.10(a), Fig.10(b), Fig.10(c), and Fig.10(d). Time series of residual temperatures between 1990 and 1997 over Fort Collins, CO (41oN, 105oW): (a) at 86 km, (b) at 100 km, along with the associated least squares fits, (c) and (d) 40-day averaged residual temperatures at 86 and 100 km and the least squares fit to raw data points as in (a) and (b).  The time scale is in year, starting January 1, 1990 as zero. The arrows mark the time of Pinatubo eruption.

Fig.11(a) and Fig.11(b). Time series of nightly mean temperatures between 1990 and March 1999 for (a) 86 km and (b) 100 km. Notice that solar minimum occurred in June 1996.

Fig.12(a) and Fig.12(b).  Left, temperature change between solar maximum and solar minimum, observed (dot-dash) and simulated (dot), along with the rate of non-solar temperature change (solid). Right, enhanced solar response observed in winters and QBO-West phase (from Fig. 5a of Chanin et al., 1989).
 

References:

Agnelli, G., A. Cacciani, and M. Fofi, 1975: The magneto-optic filter, Solar Phys. 44, 509.

Bills, R. E., and C. S. Gardner, 1993: Lidar observations of the mesopause region temperature structure at Urbana, J.
Geophys. Res. 98, 1011-1021.

Bills, R. E., C. S. Gardner, and S. J. Franke, 1991: Na Doppler/temperature lidar: initial mesopause region observations and
comparison with the Urbana medium frequency radar, J. Geophys. Res. 96, 22,701-22,707.

Brasseur, G. and M. H. Hitchman, Stratospheric response to trace gas perturbations: changes in ozone and temperature
distribution, Science, 240, 634-637, 1988.

Chen, H., M. A. White, D. A. Krueger, and C. Y. She, 1996: Daytime mesopause temperature measurements using a
sodium-vapor dispersive Faraday filter in lidar receiver, Opt. Lett. 21, 1093-1095.

Chen, S. S., Z. L. Hu, M. A. White, D. A. Krueger and C. Y. She, Lidar observations of seasonal variation of diurnal mean
temperature in the mesopause region over Fort Collins, CO (41oN, 105oW), J. Geophys. Res., 2000.

Chanin, M.-L., P. Keckhut, A. Hauchecorne, and K. Labitzke, The solar activity ?Q.B.O. effect in the lower thermosphere,
Ann. Geophysicae, 7, 463-470, 1989.

Fleming, E. L., S. Chandra, J.J. Barnett, and M. Corney, 1990: Zonal mean temperature, pressure, zonal wind, and
geopotenyial height as functions of latitude, Adv. Space Res. 10, 11-59.

Fricke, K.H., and U. von Zahn, 1985: Mesopause temperatures derived from probing the hyperfine structure of the D2
resonance line of sodium by lidar, J. Atmos. Terres. Phys. 47, 499-512.

Keckhut, P., A. Hauchecorne, and M. L. Chanin, Midlatitude long-term variability of the atmosphere: Trends and cyclic and
episodic changes, J. Geophys. Res., 100, 18,887-18,897, 1995.

Hair, J. W., David A. Krueger, and C. Y. She, High Spectral resolution lidar at 532 nm using iodine vapor filters: a working
proto-type for simultaneous measurement of atmospheric state parameter and aerosol parameters, submitted (2000)

Hauchecorne, A. and M-L Chanin, Density and temperature profiles obtained by lidar between 35 and 70 km, Geophys. Res.
Lett., 78, 565-568, 1980.

Huang, T. Y. W. and G. P. Brasseur, Effect of Long-term solar variability in a two-dimensional interactive model of the middle
atmosphere, J. Geophys. Res., 98, 20,413 ?20,427, 1993.

Jeys, T. H., A. A. Brailove, and A. Mooradian, 1989: Sum frequency generation of sodium resonance radiation, Appl. Opt 28,
2588-2591.

Krueger, D. A., H. Chen, and C. Y. She, 1997: Mesopause temperature and sodium density variations over 35 hours during
Oct. 1995, at 40.59oN, 105.14oW, Proc. 13th ESA Symposium on European Rocket and Balloon Programmes and Related
Research, Oland, Sweden.

Krueger, D. A., and C. Y. She, Observed “long-term?temperature change in a midlatitude mesopause region in response to
external perturbation, Earth, Planets and Space, 51, 809-814, 1999.

Moosmüller, H., and J. D. Vance, 1997: Sum-frequency generation of continuous-wave sodium D2 resonance radiation, Opt.
Lett. 22, 1135-1137.

Rind, D., R. Suozzo, N. K. Balachanaran, and M. J. Prather, Climate Changes and the middle atmosphere: Part 1: The
doubled CO2 climate, J. Atmos. Sci., 47, 475-499, 1990.

Roble, R. G., and R. F. Dickinson, How will changes in carbon dioxide and methane modify the mean structure of the
mesosphere and thermosphere ?, Geophys. Res. Lett., 16, 1441-1444, 1989.

Senft, D. C., G. C. Papen, C. S. Gardner, J. R. Yu, D. A. Krueger, and C. Y. She, 1994: Seasonal variations of the thermal
structure of the mesopause region at Urbana, IL (40oN, 88oW) and Ft. Collins, CO (41oN, 105oW), Geophys. Res. Lett.
'21, 821-824.

She, C. Y., H. Latifi, J. R. Yu, R. J. Alvarez II, R. E. Bills, and C. S. Gardner, 1990: Two-frequency lidar technique for
mesospheric Na temperature measurements, Geophys. Res. Lett. 17, 929-932.

She, C. Y., J. R. Yu, and H. Chen, 1993: Observed thermal structure of a midlatitude mesopause, Geophys. Res. Lett. 20,
567-570.

She, C. Y., and J. R. Yu, Simultaneous three-frequency Na lidar measurements of radial wind and temperature in the
mesopause region, Geophys. Res. Lett., 21, 1771-1774, 1994.

She, C. Y., and J. R. Yu, 1995: Doppler-Free Saturation Fluorescence Spectroscopy of Na Atoms for Atmospheric
Applications, Appl. Opt., 34, 1063-1075.

She, C. Y., J. R. Yu, D. A. Krueger, R. Roble, P. Keckhut, A. Hauchecorne, and M.-L. Chanin, 1995: Vertical structure of
the midlatitude temperature from stratosphere to mesosphere (30-105 km), Geophys. Res. Lett. 22, 377-380.

She, C. Y., and U. von Zahn, 1998: The concept of two-level mesopause: Support through new lidar observation, J. Geophys.
Res. 103, 5855-5863.

She, C. Y., S. W. Thiel and D. A. Krueger, Observed episodic warming at 86 and 100 km between 1990 and 1997: Effects
of Mount Pinatubo eruption, Geophys. Res. Lett., 25, 497 - 500, 1998.

She, C. Y., Spectral Structure of Laser Light scattering Revisited: Bandwidths of Non-resonant Scattering Lidars, submitted
(2000).

States, R. J., and C. S. Gardner, Temperature structure of the mesopause region (80-105 km) at 40oN latitude, 1, Seasonal
variations, J. Atmos. Sci., 57, 66-77, 2000.

Vance, J. D., C. Y. She,and H. Moosmüller, 1998: Continuous-wave, all-solid-state, single-frequency 400-mW source at 589
nm based on doubly resonant sum-frequency mixing in a monolithic lithium-niobate resonator, Appl. Opt. 37, 4,891-4,896.

von Zahn, U., J. Hoffner, V. Eska, and M. Alpers, 1996: The mesopause altitude: Only two distinctive levels worldwide?,
Geophys. Res. Lett. 23, 3231-3235.

Williams, B. P., C. Y. She, and R. G. Roble, Seasonal climatology of the nighttime tidal perturbation of temperature in the
midlatitude mesopause region, Geophys. Res. Lett., 25, 3301 - 3304, 1998.

Yeh, P., 1982: Dispersive magneto-optic filter, Appl. Opt 21, 2069.

Yu, J. R., and C. Y. She, 1995: Climatology of a midlatitude mesopause region observed by a lidar at Fort Collins, Colorado
(40.6oN, 105oW), J. Geophys. Res. 100, 7441-7452.