A comparison between ground-based observations of noctilucent clouds and Aura satellite data

Journal of Atmospheric and Solar-Terrestrial Physics 73 (2011) 2097–2109

Contents lists available at ScienceDirect

Journal of Atmospheric and Solar-Terrestrial Physics journal homepage: www.elsevier.com/locate/jastp

A comparison between ground-based observations of noctilucent clouds and Aura satellite data P. Dalin a,b,n, N. Pertsev c, A. Dubietis d, M. Zalcik e, A. Zadorozhny f, M. Connors g, I. Schofield g, T. McEwan h, I. McEachran h, S. Frandsen i, O. Hansen j, H. Andersen j, V. Sukhodoev c, V. Perminov c, R. Balcˇiunas d, V. Romejko k a

Swedish Institute of Space Physics, Box 812, SE-981 28 Kiruna, Sweden Space Research Institute, RAS, Profsouznaya st. 84/32, Moscow 117997, Russia c A. M. Obukhov Institute of Atmospheric Physics, RAS, Pyzhevskiy per., 3, Moscow 119017, Russia d Department of Quantum Electronics, Vilnius University, Saule_ tekio Ave. 9, bldg. 3, LT-10222 Vilnius, Lithuania e NLC CAN AM Network, #7 14130 80 Street, Edmonton, AB, Canada T5C 1L6 f Novosibirsk State University, Pirogova st. 2, Novosibirsk 630090, Russia g Athabasca University Geophysical Observatory, Athabasca, Alberta, Canada T9S 3A3 h NLC NET, 14 Kersland Road, Glengarnock, Ayrshire, KA14 3BA Scotland, UK i University of Aarhus, Ny Munkegade, Bygn. 1520, DK-8000 Aarhus C, Denmark j The Danish Association for NLC Research, Lyngvej 36, Kolvr˚ a, DK-7470 Karup J., Denmark k The Moscow Association for NLC Research, Kosygina st. 17, 119334 Moscow, Russia b

a r t i c l e i n f o

abstract

Article history: Received 1 April 2010 Received in revised form 13 January 2011 Accepted 27 January 2011 Available online 4 February 2011

A comparison is made between ground-based observations of noctilucent clouds (NLCs), obtained with a network of automated digital cameras, and Aura satellite data (the MLS instrument). The Aura data (water vapor and temperature) demonstrate reasonable values around the summer mesopause fostering NLC formation in June through August, when supersaturated air conditions occur. The temperature decrease leads, in general, to amplification of the NLC brightness. The 2- and 5-day planetary waves, extracted from the Aura temperature field, have definite influence on the brightness variations of NLCs. The temperature behavior around the summer mesopause at 601N demonstrated a remarkable feature, namely, in 2007 the minimum of a Gaussian fitted seasonal temperature variation was observed, on average, 14 days earlier and was broader than the corresponding minimum in 2008. The different temperature climatology resulted in different seasonal variation of NLCs in 2007 and 2008; in particular, the maximum of a Gaussian fitted seasonal variation of the NLC brightness cycle in 2007 was advanced by 12–29 days relative to that in 2008. & 2011 Elsevier Ltd. All rights reserved.

Keywords: Noctilucent clouds Planetary waves Atmospheric dynamics

1. Introduction The highest clouds in the Earth’s atmosphere are noctilucent clouds (NLCs) formed around the mesopause in the 80–90 km altitude range. These clouds are a beautiful nighttime optical phenomenon occurring during the summer months at mid- and high latitudes. NLCs consist of water ice particles of 30–100 nm in radius that scatter sunlight and thus NLCs are readily seen against the dark twilight arc from late of May until September (Gadsden ¨ and Schroder, 1989). Satellite observations have discovered polar mesospheric clouds (PMCs) covering the entire polar mesopause region above

n Corresponding author at: Swedish Institute of Space Physics, Box 812, SE-981 28 Kiruna, Sweden. Tel.: + 46 980 79023; fax: + 46 980 79050. E-mail address: [email protected] (P. Dalin).

1364-6826/$ - see front matter & 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.jastp.2011.01.020

a 701 latitude during summertime (Donahue et al., 1972). Sometimes, some ‘‘patches’’ of PMCs (similar to icebergs from a continent) extend to mid-latitudes and become visible from the Earth’s surface as NLCs. But it is poorly understood so far what characteristic sizes of such patches are and which processes are responsible for their formation. Planetary atmospheric waves are one of the major players in the middle atmosphere producing strong disturbances of the basic state of the atmosphere. The temperature disturbances of 1–8 K due to planetary waves (Espy and Witt, 1996; Pogoreltsev, 1999) yield a correlation between the probability of NLC appearance and the combined effect of stationary, 16- and 5-day planetary waves (Kirkwood and Stebel, 2003). The 5-day period in NLCs has been detected using ground-based observations from Western Europe (Gadsden, 1985; Kirkwood and Stebel, 2003). Dalin et al. (2008) have demonstrated a clear influence of the 2- and 5-day planetary waves on the occurrence frequency,

2098

P. Dalin et al. / Journal of Atmospheric and Solar-Terrestrial Physics 73 (2011) 2097–2109

geographical distribution and brightness variations of NLCs observed with a network of automatic cameras located along a 55–561 latitude circle around the globe. The signatures of the 2- and 5-day planetary waves were also observed in PMCs from satellite data (Merkel et al., 2003, 2008). The authors have used the SNOE satellite and TIMED (SABER instrument) spacecraft data to observe PMCs at high latitudes 68–801 and retrieve temperature variations, respectively. Merkel et al. (2008) have found that during the 2002 and 2003 summer mesosphere seasons the temperature perturbations due to 2- and 5-day waves were small (2.0–3.5 K), but had a significant effect on the PMC brightness variation. Von Savigny et al. (2007) have reported on simultaneous measurements of the quasi 5-day wave in NLCs and the mesopause temperature. Their analysis has employed the Envisat satellite limb scattering measurements to detect NLCs/PMCs in the latitude band between 601 and 801 and the Aura satellite data to measure the mesopause temperature. They have found that for some periods (before and around the solstice) the quasi 5-day wave signatures in NLCs were likely caused by quasi 5-day wave signatures in the temperature field. Also, von Savigny et al. (2007) have noted high variability of NLCs (2–3 days) for certain intervals of the 2005 NLC season and questioned whether this variability may be related to the 2-day planetary wave. In the present paper, for the first time, we compared groundbased observations of NLCs with the hemispheric network of automatic cameras and the Aura (the MLS instrument) satellite data to obtain information on humidity and temperature regime as well as to retrieve the planetary wave propagation around the mesopause in summer of 2007 and 2008.

2. Data source After several years of experience in Moscow and Novosibirsk (Russia), an international network started to work in 2004 in Moscow, Novosibirsk and Lund (Sweden) to continuously monitor NLCs. This network has been extended for successive years, and at present it includes cameras situated in Port Glasgow, Scotland (551N560 ; 041W410 ), Athabasca, Canada (541N440 ; 1131W190 ), Kamchatka, Russia (531N040 ; 1581E370 ), Novosibirsk, Russia (541N520 ; 831E060 ), Moscow, Russia (561N000 ; 371E290 ), Vilnius, Salakas and Vidiskes, Lithuania (average coordinates 551N000 ; 261E000 ), Aarhus, Denmark (561N100 ; 101E120 ) in 2007 and 2008. Each camera operates from the end of May until middle of August and takes images every 1 min at night during high NLC season (June 10–July 25). The detailed description of the network of NLC cameras and operation schedule can be found in Dalin et al. (2008). The most important feature is that these NLC cameras are located along approximately the same latitude circle of 55–561. Such geographical camera locations provide comparable NLC observations (due to equal twilight illumination conditions, as well as very similar physical conditions in the mesopause since temperature, vertical and horizontal winds are latitude dependent), and provide us with the possibility to study the NLC activity on continental scales, including effects due to gravity and planetary waves. The following characteristics of NLCs are included into the resulting statistics for a given season: time of appearance and disappearance of NLCs, brightness, morphological forms of NLCs, weather conditions. The NLC brightness is visually estimated on a traditional 5-point scale. Detailed description of estimating the NLC parameters can be found in Romejko et al. (2003). Note that weather conditions are of importance for the NLC statistics, and we carefully select nights with good and bad weather in order to determine a real absence of NLCs on a given night.

Another data source is the Microwave Limb Sounder (MLS) instrument onboard the NASA Aura satellite which is currently operating (Waters et al., 2006). On 15 July 2004, Aura was launched into a near-polar (981 inclination), sun-synchronous orbit 705 km above the Earth, with ascending and descending equator-crossing times at 13:45 and 01:45 local time, respectively. The orbital period is about 100 min. Since the Aura orbit is nearly fixed at solar local time, Aura makes measurements for a given point in the atmosphere twice a day (ascending and descending transversals); the number of orbits per day is around 14.6. The temperature and water vapor data of the MLS instrument from 1 June to 15 August 2007 and 2008 are used in this study. The description on the MLS temperature product and its validation can be found in Froidevaux et al. (2006) and Schwartz et al. (2008). The validation of water vapor data is described in detail by Read et al. (2007) and Lambert et al. (2007). The Aura/MLS data ver.2.2 were obtained from the NASA public web-site: http://disk.sci. gsfc.nasa.gov/Aura/data-holdings/MLS/ml2t.002.shtml.

3. Methodology of the data analysis It is important to note that the Aura data (and data of any satellite) are discrete and asynchronous in nature, the universal time of an observation and longitude is mutually dependent; satellite’s orbital tilt results in irregular sampling between ascending and descending nodes (Salby, 1982). Also, since the number of orbits per day is non-integer, there is a phase drift of the longitudinal nodes for any particular latitude circle for subsequent days. Besides, data voids usually exist. To overcome many of these problems one can make an interpolation of asynoptic satellite observations in space (in longitude and latitude) and time on a fixed grid, retrieving a twice-daily sequence of synoptic maps (Rodgers, 1976); this technique is employed in the present paper. We apply a double Fourier transform (in longitude and time) to the temperature field for each latitude (horizontal spacing in latitude is 1.51) from  501 to + 681 and height from 20 to 110 km. Riggin et al. (2006) have employed the same procedure in the analysis of the SABER data onboard the TIMED satellite. The time resolution of a sequence of synoptic maps is 12 h. The vertical grid is defined on 29 levels, with the vertical resolution varying from 3 km in the stratosphere to 12–14 km around the mesopause. Since 15 longitudes are sampled twice a day, the periodic oscillations with zonal wavenumbers from  7 to 7 are possible to retrieve, with the Nyquist frequency being 1 day  1. The time series of temperature amplitude and phase (expressed as a complex amplitude for westward propagating zonal wavenumber one, W1) is filtered by a 5-pole bi-directional Butterworth filter, in the frequency domain that passes periods between 4 and 7 days, to retrieve the 5-day component. By using a filter with complex coefficients, there can be extracted westward traveling waves only; by employing a bi-directional filter, the artificial phase shifts are eliminated. In the same manner, temperature perturbations due to a westward traveling 2-day planetary wave of zonal wavenumber 3 (W3) have been obtained by applying a band-pass filter to reveal oscillations with periods between 1.7 and 3.2 days. The planetary wave analysis has been performed on both the nighttime and daytime data to get as high temporal resolution as possible (12 h) to avoid a possible aliasing effect in the Fourier analysis for the 2-day planetary wave. Thus, time series of temperature field perturbations due to the westward propagating 2- and 5-day planetary wave, as a function of height and latitude, are derived. Then, only periodic oscillations along a 601N latitude circle are chosen to be compared with observations of NLCs. Estimations of the NLC brightness are compared to temperature disturbances, and finally the linear regression analysis is employed

P. Dalin et al. / Journal of Atmospheric and Solar-Terrestrial Physics 73 (2011) 2097–2109

2099

and 301, that is from 71 to 1.31 north of the observer (Fogle and Haurwitz, 1966). The latitude of 601N is a round number which is nearest to the apparent geometrical center of the NLC field as seen from 551 to 561. Therefore the MLS temperature data are considered for 601N in the present paper.

to make conclusions on the degree of influence of the 2- and 5-day planetary wave on the NLC activity. The latitude of 601N has been chosen due to the following reason. It is a well established fact that vast majority of NLCs observed from the ground are seen at elevation angles between 31

Port Glasgow 2007

160 T [K]

1

4

150 2

140

3

3

130 150

2

1

3 1

160

2 3 2 2.5 1 3.5 4 2 .5 3.53 3 2 4 0.5 3 2.5

170

180

2

2

3 3

190

2.5

4

3

200

210

220

230

Athabasca 2007

T [K]

160 150 1 1.5 2

140 130 150

2.52 3

1

160

2

3.5

0.5 2.5

2.5 1

43.5

170

180

190

200

1.5

210

220

230

210

220

230

Novosibirsk 2007

T [K]

160 150 2

140

3

1.5 3

130 150

3

2.5 3 1.5 2 3

1

160

4 2.52

0.5

4.5 3 1

170

1

21

2

180

5 4

190

4

200

2.5

Moscow 2007

T [K]

160 150 140

0.5

0.5

4 41 3.5 3 22 2.5 4

3

130 150

160

0.5 4.522.5 2 2 4 1.5

170

4

180

1.5

1 1 1.5 3.52.5

5

1

2

190

200

210

220

230

220

230

Vilnius 2007

T [K]

160 150 140 130 150

2

1

12

160

42 2

3.5 2

170

4

2 4 2

42

180

2

3 1.5

2

2

2

2.5

190 Aarhus 2007

200

210

160 T [K]

150

5

140

4

1

130 150

160

170

5 5

3

180

2

3

190

2.5

4

2

200

2.5

2.5

5

210

220

230

DoY, 2007 Fig. 1. The 2007 seasonal behavior of NLC occurrences and NLC brightness as well as the MLS mesopause temperature (black line) and frost point temperature (blue line) at 0.0046 mb or 85 km at 601N. Nights with NLCs are indicated by scaled points with appropriate digits of the NLC brightness. The red curve is a Gaussian fit in the leastsquare sense to the MLS mesopause temperature. Gaps are due to the Aura data voids on day 194, 219 and 220 of year. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

2100

P. Dalin et al. / Journal of Atmospheric and Solar-Terrestrial Physics 73 (2011) 2097–2109

purposes of validation of the Aura temperature and water vapor values around the summer mesopause.

Also, the MLS water vapor as well as temperature data sets have been directly used (without any interpolating and filtering procedure) to calculate the frost point temperature around the mesopause, at longitudes closest to each observational site and at a 601 latitude circle. The estimation of the frost point temperature ¨ is based on a formula presented by Gadsden and Schroder (1989). The time series of the mesopause and frost point temperature is directly compared to the NLC brightness variations for each observational site. Such a direct comparison provides knowledge on the NLC climatology as well as valuable information for

4. Data analysis Before we will consider the comparison between the MLS data and NLC observations, it is necessary to mention the following. According to Schwartz et al. (2008), the horizontal resolution of the MLS temperature measurements in the mesosphere is

Port Glasgow 2008

T [K]

160 150

3

140

2

130 150

4 1

2.53 2 3 1.5

0.5

160

3 3 4

1

170

180

3

2

3.5

243

190

0.5

2.5 0.5

1 4 2.5

200

4

210

220

230

220

230

Athabasca 2008

T [K]

160 150 140

2

130 150

0.5 1 5 4.5 0.5 1 1 3 2 3

1 1.5

3

3

160

170

180

3 4

190

1

1

1

2 3.5

200

210

Novosibirsk 2008

T [K]

160 150 140 130 150

1.5

1

4

2.5 5 2.5

1

160

170

2 253

180

4 2

2 2 2.5 3

190

1.53

2

2.5

1

200

210

2

0.5

220

230

Moscow 2008

T [K]

160 150 140

2.5

2 0.5

1

130 150

160

2

3.53.5 1.5 1 2.5 2 4

170

180

1.5 3 3 4.5 3.5

190

1

1 1 2 2.5

1

2

2

3

200

210

220

230

220

230

Vilnius 2008

T [K]

160 150 140

2.5

1 2 2.5

5 1.5 1 3

170

180

130 150

160

3.5 2.5 4 4.5 1.52 3.5 4 3 2.5

190

1 43

4

2

1 4

200

1 3 0.5 3 0.5 1.5 1.5

210

Aarhus 2008

T [K]

160 150 4

140

2

130 150

170

0.5 51

0.5

160

4.5 1 1

180

1.5 3 2 0.5

190

1 2

2.5 2 3 4.5 1 4

200

2 2 1.5 0.5 1 3.5

210

2

2

3

220

230

DoY, 2008 Fig. 2. The 2008 seasonal behavior of NLC occurrences and NLC brightness as well as the MLS mesopause temperature (black line) and frost point temperature (blue line) at 0.0046 mb or 85 km at 601N. Nights with NLCs are indicated by scaled points with appropriate digits of the NLC brightness. The red curve is a Gaussian fit in the leastsquare sense to the MLS mesopause temperature. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

P. Dalin et al. / Journal of Atmospheric and Solar-Terrestrial Physics 73 (2011) 2097–2109

200 km along the line-of-sight (LOS) by 12 km across the LOS. At the same time, Aura orbit tracks are separated by  1400 km at 601 latitude that means the nearest MLS measurement could be about 700 km east or west of an observational point in a worse case. On the other hand, it is also true that the NLC field usually extends over several hundred km along a latitude circle with the characteristic scale of less than 800 km (Dalin et al., 2006). Since the horizontal field of view of the camera is limited to around 601 (56–631, depending on camera and lens model), only a part of a real NLC field (which in fact is larger) is captured as it travels through the camera field of view. Thus, even if the nearest MLS measurement could be about 700 km east or west of the observer, there is a large probability that the MLS measurement fall into the NLC field due to a large extent of the last one. Such spatial differences (of several hundreds km) are significantly less than wavelengths of planetary waves considered in the present work. It is necessary to point out another very important issue when considering a comparison between NLC ground-based observations and satellite measurements. The summer mesopause represents a highly dynamical region of the atmosphere. Indeed, we always observe gravity waves of different scales in NLCs. Larger distances do not necessarily mean larger uncertainties (errors) in temperature because NLCs may readily lie on a successive crest of medium- or large-scale gravity waves (more than 100 km) with the same temperature disturbances, which are rather large around the summer mesopause ( 75–10 K, Rapp et al., 2002). The Aura measurement is made at 02:30–02:40 solar local time at 601N on a descending node while NLCs are typically observed between 23:30 and 04:00 Daylight Saving Time (Time Zone Local Time (LT)+ 1 h). Thus, taken Aura measurements are within the time interval of the NLC observation. Temporal difference of 2–3 h is possible only in the case of short living NLCs. Nevertheless, this time difference is short compared to long period atmospheric processes regarded. Since we choose Aura measurements at the fixed solar local time, the non-migrating diurnal temperature variations are eliminated. If we take temporal difference of 2–3 h, the temperature difference due to migrating diurnal tidal amplitudes is expected to be 1.5–2.0 K, according to assimilation model results (including the MLS temperature measurements) of Eckermann et al. (2009), who found migrating diurnal temperature variations of 6–8 K at 601N at the NLC height. Such temperature variations of 1.5–2.0 K (if exist) provide rather small source of uncertainty in the analysis of an individual NLC observation compared to uncertainties of 5–10 K due to gravity waves.

2101

4.1. Comparison of the MLS temperature behavior with NLC observations For 2007, NLC occurrences and NLC brightness along with the MLS mesopause temperature (at 0.0046 mb or 85 km) as well as the frost point temperature for appropriate longitudes is shown in Fig. 1. The data for Port Glasgow, Athabasca, Novosibirsk, Moscow, Vilnius and Aarhus are illustrated in corresponding subpanels. The mesopause temperature varies around 145–150 K in the beginning of June, then drops to 130–135 K between 10 and 30 June (around the solstice) and successively increases to 160–165 K in the middle of August. The amount of water vapor varies between 3.5 and 6.0 ppmv, yielding variation of the frost point temperature in the range of 145–147 K during the whole season. In general, nights with NLCs (indicated by scaled circles with an appropriate digit of the NLC brightness) are shown to appear from the beginning of June until the middle of August. This is a typical seasonal duration which agrees well with previous obser¨ vational NLC studies (e.g., Gadsden and Schroder, 1989). The NLC occurrence frequency is rather low and NLC brightness is faint or moderate from 1 to 10 June and from 1 to 15 August; moderate and bright NLC is frequently observed from about 15 June to 1 August that coincides with the general temperature decreasing from lowest to moderate low values between 130 and 145 K. On

Table 2 Statistical characteristics of a Gaussian distribution fitted to the NLC brightness variations for 2007 and 2008. Observational sites in 2007/2008

Mean day (DoY)

FWHM, day 1 and day 2

Port Glasgow 07/08

174.9/190.9 Diff¼ 16.0 175.0/187.0 Diff¼ 12.0 166.9/196.0 Diff¼ 29.1 175.7/190.5 Diff¼ 14.8 181.2/194.0 Diff¼ 12.8 180.0/201.5 Diff¼ 21.5 175.6/193.3 Diff¼ 17.7

151.1–197.7/165.0–216.8

Athabasca 07/08 Novosibirsk 07/08 Moscow 07/08 Vilnius 07/08 Aarhus 07/08 All sites

163.7–186.3/179.0–195.0 134.1–199.7/179.7–212.2 158.4–193.0/172.7–208.4 168.0–194.4/177.4–210.5 174.2–185.8/180.4–222.6 158.3–192.8/175.7–210.9

Table 1 Statistical characteristics of a Gaussian distribution fitted to the MLS temperature behavior for 2007 and 2008. FWHM stands for full width at half maximum. Day 1 and day 2 stand for the width of a Gaussian fit. Observational sites in 2007/2008

Mean day (DoY)

FWHM, day 1 and day 2

Minimum T (K)

Port Glasgow 07/08

178.0/182.7 Diff ¼4.7 168.1/181.6 Diff ¼13.5 161.3/181.8 Diff ¼20.5 174.1/184.0 Diff ¼9.9 166.7/183.3 Diff ¼16.6 176.5/181.6 Diff ¼5.1 170.8/182.5 Diff ¼11.7

143.6–212.5/151.0–214.4

137.7/137.8

126.6–209.6/149.1–214.0

139.7/139.7

115.1–207.7/149.4–214.2

139.6/138.1

137.0–211.3/153.2–214.8

138.5/136.7

124.3–209.1/152.1–214.6

139.4/138.8

141.0–212.0/148.4–214.8

138.2/138.0

131.3–210.4/150.3–214.5

138.9/138.2

Athabasca 07/08 Novosibirsk 07/08 Moscow 07/08 Vilnius 07/08 Aarhus 07/08 All sites

2102

P. Dalin et al. / Journal of Atmospheric and Solar-Terrestrial Physics 73 (2011) 2097–2109

the other hand, there are a number of NLC cases as seen from Port Glasgow, Novosibirsk, Moscow, Vilnius and Aarhus when the mesopause temperature is above the frost point temperature. This fact has also been noted by Dubietis et al. (2010) in the analysis of the Lithuanian NLC observations and Aura/MLS data. We provide explanations for this result in the Discussion section. The comparison between the NLC brightness and MLS temperature for 2008 is presented in Fig. 2. The overall picture is similar to that of 2007: the majority of NLC start to appear when mesopause temperatures fall below the frost point temperature, and on several occasions NLCs were observed from Port Glasgow, Moscow, Vilnius and Aarhus under unsaturated air conditions. At the same time, one can note a remarkable difference in the temperature behavior in 2007 and 2008: the minimum temperatures are observed from 161 to 178 days in 2007, and from 182 to 184 days in 2008, and the apparent time difference is about 10 days. To quantitatively verify these visual graphical estimations, Gaussian fits in the least-square sense to the mesopause temperature variations as well as to the NLC brightness data of 2007 and 2008 have been constructed; note that zero NLC brightness (absence of NLCs on a semi-clear or clear night) has been included into the statistics. The statistical characteristics of Gaussian curves for the temperature and NLC season are given in Tables 1 and 2, respectively; a Gaussian distribution for the NLC brightness is not shown in Figs. 1 and 2. A time difference between the statistical minima (statistical minimum is defined as a temperature minimum of a Gaussian fit) of the temperature behavior of 2007 and 2008 is evident from Table 1: for Port Glasgow is 4.7 days, for Athabasca is 13.5 days, for Novosibirsk is 20.5 days, for Moscow is 9.9 days, for Vilnius is 16.6 days and for Aarhus is 5.1 days. Also note that the width of Gaussian curves for 2007 is systematically broader by 5–28 days (depending on a site) compared to one for 2008. Table 2 demonstrates similar time differences between the statistical maxima of the NLC brightness cycle of 2007 and 2008:

for Port Glasgow it is 16 days, for Athabasca 12 days, for Novosibirsk 29.1 days, for Moscow 14.8 days, for Vilnius 12.8 days, for Aarhus 21.5 days. It is clear that some prominent event took place in the inter-annual temperature climatology at the mesopause, resulted in great time differences of the NLC climatology between 2007 and 2008. It is worth to consider not only a general link between a seasonal drop of the summer mesopause temperature and NLC cycle, but also a relationship between the changes in temperature and NLC brightness in detail. To characterize this relationship, we have applied the linear regression analysis for the data at each observational point (not shown in figure). This analysis demonstrates distinct negative trends for all the sites for both years meaning that temperature decrease leads to amplification of the NLC brightness. For 2007, the trends are significant for all the observational points except of Aarhus; for 2008, the trends are significant for Athabasca, Novosibirsk, Moscow and Vilnius, and are insignificant for Port Glasgow and Aarhus. To increase the statistical significance, the data of all the sites have been combined together, and regression analysis applied is presented in Fig. 3. The negative trends are highly significant for both years,

Table 3 The correlation coefficient (r) and the square of the coefficient of determination (R2) between the Aura/MLS temperature variations and NLC brightness for 2007 and 2008. Site

r

R2

Port Glasgow 07/08 Athabasca 07/08 Novosibirsk 07/08 Moscow 07/08 Vilnius 07/08 Aarhus 07/08 All sites 07/08

 0.45/  0.16  0.47/  0.55  0.37/  0.51  0.37/  0.40  0.34/  0.54  0.22/  0.13  0.38/  0.39

0.20/0.03 0.22/0.30 0.14/0.26 0.13/0.16 0.12/0.29 0.05/0.02 0.14/0.15

All data 2007 165

y= –1.68x + 145.0

160 T[K]

155 150 145 140 135 130 0

0.5

1

1.5

2

2.5

3

3.5

4

4.5

5

All data 2008 155

y= –1.58x + 143.7

T[K]

150 145 140 135 130 125 0

0.5

1

1.5

2 2.5 3 3.5 NLC brightness [points]

4

4.5

5

Fig. 3. Linear regression analysis for the Aura/MLS temperature versus NLC brightness: the upper panel is for 2007, the lower panel is for 2008. The data of all six observational sites are combined into a single data set for each year.

P. Dalin et al. / Journal of Atmospheric and Solar-Terrestrial Physics 73 (2011) 2097–2109

2103

Fig. 4. Panel a: the 5-day wave pattern as a function of longitude and time for 2007 from Aura MLS temperatures at 0.0046 mb (85 km). Panel b: the 5-day wave pattern as a function of height and time for 2007 at 121E. The Local Solar Time is  02:30 and  12:00.

Fig. 5. Panel a: the 5-day wave pattern as a function of longitude and time for 2008 from Aura MLS temperatures at 0.0046 mb (85 km). Panel b: the 5-day wave pattern as a function of height and time for 2008 at 1081W. The Local Solar Time is  02:30 and  12:00.

indicating that the brightness of noctilucent clouds increases with decreasing of temperature. This result highlights, on the one hand, intimate relationship between temperature and NLC brightness, and, on the other hand, the reliability of the NLC brightness and temperature estimations. It is useful to estimate the statistical contribution of the variance due to temperature variations to the total variance of the NLC brightness. It is expressed as the square of the coefficient of determination which in turn equals to the square of the correlation coefficient between the observed and predicted data

values when considering the linear regression model. Values of the correlation coefficient and the square of the determination coefficient for 2007 and 2008 are presented in Table 3. It is seen that the temperature variance explains the total variance of the NLC brightness by 5–22% for 2007 and 2–30% for 2008, depending on data set, and the average value of R2 is equal to 14% and 15% for 2007 and 2008, respectively. As will be shown below, temperature variations contribute to the total variance of the NLC brightness to a greater extent than temperature disturbances due to 2- and 5-day planetary waves.

2104

P. Dalin et al. / Journal of Atmospheric and Solar-Terrestrial Physics 73 (2011) 2097–2109

4.2. The 5-day planetary wave analysis and NLCs The examples of the 5-day wave patterns in the longitude– time and height–time cross-section for 2007 are illustrated in the upper and lower panel of Fig. 4, respectively. The longitude–time dependence is given for the height of the mesopause and for a 601 latitude circle. Similar 5-day wave patterns for 2008 are demonstrated in Fig. 5. Clear westward propagating phase structures are seen in Figs. 4a and 5a, which have been shown and described in many studies (for example, Salby, 1984). Temperature disturbances for this period are up to 73 K. Figs. 4b and 5b show vertically propagating phase structures for the longitudes close to Aarhus and Athabasca, respectively (wave patterns for other longitudes look rather similar). The maximum amplitudes up to 73 K are observed in the range of 80–110 km. Note that the activity of the 5-day wave varies in time, sometimes it vanishes and almost disappears and sometimes grows again. Now we aim to study the effect of the extracted 5-day waves on direct observations of NLCs around the globe. Probably the best way is to compare ‘‘raw’’ NLC brightness variations with temperature ones by the wave, because, unfortunately, one cannot apply a harmonic analysis to the NLC time series which include gaps due to bad weather. The linear regression analysis has been performed for the 5-day temperature disturbances and NLC brightness for each observational point (not shown in figure). For 2007, the brightness of NLCs increases with negative temperature disturbances for each site; however, the linear regression fit is significant for the Aarhus data only and has no statistical significance for other sites. That is why we have combined the data of each site into a single data set and have applied the regression analysis which is demonstrated in Fig. 6a. Clear significant negative trend with the slope of  0.1470.04 K/Brightness Unit (BU) is obtained, indicating that the brightness of noctilucent clouds increases when temperature disturbances go to negative values.

For 2008, the linear regression analysis revealed almost identical behavior of the NLC brightness. At the same time, the linear regression fit has been found to be significant for the Athabasca data only. The regression analysis of the combined data for 2008 is shown in Fig. 6b, which demonstrates significant negative linear trend with the slope equal to 0.1170.04 K/BU, meaning that negative temperature variations lead to amplification of the NLC brightness. The difference between the slopes (0.14 and 0.11) is statistically indiscernible. The statistical contribution of the variance due to 5-day wave temperature variations to the total variance of the NLC brightness for 2007 and 2008 is presented in Table 4. It is seen that the 5-day wave variance explains the total variance of the NLC brightness by 1–15% for 2007 and 1–18% for 2008, depending on data set, and the average value the square of the coefficient of determination (R2) is equal to 4% and 3% for 2007 and 2008, respectively.

4.3. The 2-day planetary wave analysis and NLCs The patterns of the 2-day W3 planetary wave for 2007 are illustrated in the upper and lower panel of Fig. 7, which demon-

Table 4 The correlation coefficient (r) and the square of the coefficient of determination (R2) between 5-day wave temperature perturbations and NLC brightness for 2007 and 2008. Site

r

R2

Port Glasgow 07/08 Athabasca 07/08 Novosibirsk 07/08 Moscow 07/08 Vilnius 07/08 Aarhus 07/08 All sites 07/08

 0.29/  0.25  0.08/  0.43  0.12/  0.09  0.26/  0.12  0.16/  0.07  0.39/  0.08  0.20/  0.16

0.08/0.06 0.01/0.18 0.02/0.01 0.07/0.02 0.03/0.01 0.15/0.01 0.04/0.03

All sites 2007 3

y= –0.14x + 0.17

ΔT [ K ]

2 1 0 –1 –2 –3 0

0.5

1

1.5

2

2.5 3 All sites 2008

3.5

4

4.5

5

y= –0.11x + 0.15

2

ΔT [ K ]

1 0 –1 –2 0

0.5

1

1.5

2 2.5 3 3.5 NLC brightness [points]

4

4.5

5

Fig. 6. The linear regression analysis for temperature perturbations due to a 5-day wave versus NLC brightness: the upper panel is for 2007, the lower panel is for 2008. The data of all six observational sites are combined into a single data set for each year.

P. Dalin et al. / Journal of Atmospheric and Solar-Terrestrial Physics 73 (2011) 2097–2109

2105

Fig. 7. Panel a: the 2-day W3 wave pattern as a function of longitude and time for 2007. Panel b: the 2-day W3 wave pattern as a function of height and time for 2007 at 1081W. The Local Solar Time is  02:30 and  12:00.

All sites 2007 3

y= –0.17x + 0.20

ΔT [ K ]

2 1 0 –1 –2 –3 0

0.5

1

1.5

2

2.5

3

3.5

4

4.5

5

All sites 2008 3 y= –0.11x + 0.11

ΔT [ K ]

2 1 0 –1 –2 –3 0

0.5

1

1.5

2

2.5

3

3.5

4

4.5

5

NLC brightness [points] Fig. 8. The linear regression analysis for temperature perturbations due to a 2-day wave versus NLC brightness: the upper panel is for 2007, the lower panel is for 2008. The data of all six observational sites are combined into a single data set for each year.

strate temperature amplitudes as a function of longitude and time (for the mesopause height), and of height and time, respectively; the height–time cross-section is given for the longitude close to Athabasca. The disturbances maximize up to 74 K above 90 km. The 2-day wave activity varies both in time and space, enhancing and deceasing at some times. The effect of the 2-day W3 wave on NLCs has been studied by the use of regression analysis. For 2007, the slopes, found for

temperature disturbances and NLC brightness variations, are negative for each observational site (not shown in figure). The trends have statistical significance for the Port Glasgow, Novosibirsk and Moscow data. The linear regression model, constructed for the combined data set (Fig. 8a), yields a slope of  0.1770.04 K/BU, that is negative temperature variations lead to the NLC brightness increase. For 2008, slopes have been found to be negative for all sites as well as for 2007 (not shown in

2106

P. Dalin et al. / Journal of Atmospheric and Solar-Terrestrial Physics 73 (2011) 2097–2109

figure). Fig. 8b represents the regression analysis for the combined data for all six sites, which show up the significant negative linear model with the slope equal to 0.1170.04 K/BU. Table 5 contains information on the statistical contribution of the variance due to 2-day wave temperature variations to the total variance of the NLC brightness for 2007 and 2008. One can see that the 2-day wave variance accounts for the total variance of the NLC brightness by 1–13% for 2007 and 1–5% for 2008, depending on observational site. The average value of the 2-day wave variance is 5% and 3% for 2007 and 2008, respectively.

5. Discussion For 2007 and 2008, the Aura temperature and water vapor demonstrate very reasonable values on the descending (nighttime) transversals, that is in most cases and for all the observational sites, NLC occur when mesopause temperatures fall below the frost point temperature. At the same time, there are a number of NLC cases when temperatures are above the frost point temperature. This may be readily explained by spatial variations in temperature Table 5 The correlation coefficient (r) and the square of the coefficient of determination (R2) between 2-day wave temperature perturbations and NLC brightness for 2007 and 2008. Site

r

R2

Port Glasgow 07/08 Athabasca 07/08 Novosibirsk 07/08 Moscow 07/08 Vilnius 07/08 Aarhus 07/08 All sites 07/08

 0.36/  0.20  0.20/  0.17  0.36/  0.17  0.31/  0.23  0.08/  0.09  0.16/  0.19  0.23/  0.16

0.13/0.04 0.04/0.03 0.13/0.03 0.10/0.05 0.01/0.01 0.03/0.03 0.05/0.03

between grid point temperatures since the Aura spatial resolution in longitude between successive orbits is low, about 24.71 that corresponds to 1370 km along a 601 latitude circle; it is also possible due to low height resolution of 12–14 km suggesting the smoothing of the actual temperature around the mesopause as well as due to temporal differences between NLC observations and Aura measurements in the case of short-living NLCs. Besides, we observe events when the NLC field is separated by large dark spaces which may be caused by warm phases of medium- and large-scale gravity waves. There are even events when moving NLCs meet a ‘‘barrier’’ and rapidly disappear crossing it, suggesting a warm phase of a gravity wave. Thus it is possible that a particular instantaneous MLS measurement falls into a warm phase of a large gravity wave even being in a common volume with a total NLC field, producing the observed effect (for several cases) when NLCs and the mesopause temperature is above the frost point temperature. Concerning the Aura/MLS temperature nighttime values ¨ we should note the following. Lubken (1999) estimated the thermal structure of the arctic summer mesosphere based on in-situ measurements with falling spheres for 10 years. His results show that the temperatures around the summer mesopause (85–88 km, June–July) vary between 129 and 143 K. The Aura temperatures are found to be in the range of 130–145 K at 85 km in June and July, with the statistical value of the temperature minimum has been found to be equal to 138–140 K both for 2007 and 2008. Thus, in spite of the coarse MLS resolution near the summer mesopause, Aura temperatures are the same as ones ¨ found by Lubken (1999) for the 85 km altitude at about 701N. It is surprising to observe a great difference in the climatology conditions around the mesopause for 2007 and 2008, namely, the minimum of a Gaussian fitted seasonal temperature variation of 2008 was shifted further by 5–20 days (depending on longitude) compared to that of 2007. Also, the temperature minimum of

T [K]

T [K]

Aura T zonal mean for 2007 and 2008 157 155 153 151 149 147 145 143 141 139 137 135 150

157 155 153 151 149 147 145 143 141 139 137 135 150

T zonal mean for 2007 Gaussian fit for 2007 T zonal mean for 2008 Gaussian fit for 2008

160

170 180 190 200 210 Aura T zonal mean for 2005 and 2006

220

230

220

230

T zonal mean for 2005 Gaussian fit for 2005 T zonal mean for 2006 Gaussian fit for 2006

160

170

180

190

200

210

Days of year Fig. 9. Zonal mean temperatures at 85 km at 601N for the descending (nighttime) transversals of the Aura orbit; Gaussian fits are overplotted above temperature variations. Panel a: the solid line is for 2007, the dashed line is for 2008. Panel b: the solid line is for 2005, the dashed line is for 2006.

P. Dalin et al. / Journal of Atmospheric and Solar-Terrestrial Physics 73 (2011) 2097–2109

2007 is significantly broader by 5–28 days than one of 2008. Correspondingly, the maximum of a Gaussian fitted seasonal variation of the NLC brightness cycle in 2007 occurred 12–29 days earlier. The NLC climatology confirms this effect for each of our observational site. To find out whether these temporal shifts are of local, affected by planetary waves, or of global origin, zonal mean temperatures for the mesopause at 601N have been plotted in Fig. 9a. A similar temporal shift of the temperature development between 2007 and 2008 is clearly seen; the inter-annual difference between the minima of a Gaussian fitted seasonal temperature variation is equal to 14.4 days. This global effect cannot be explained by planetary waves because these are filtered out of zonal means. The found temporal shift of 5–20 days is a highly significant change in the summer mesopause; for some individual sites it is as long as the period of a 16-day planetary wave, for example. It is interesting to investigate what year (either 2007 or 2008) had an exceptional shift. To find it out, we have additionally analyzed the MLS temperature data for the 2005 and 2006 summer seasons and plot the results in Fig. 9b. It is clearly seen that the temperature behavior is very similar for 2005, 2006 and 2008, but it is significantly different for 2007. Thus, the temporal shift of 2007 was an exceptional one regarded to other years. The observed temporal shift might be due to well-established dynamic coupling between the stratosphere and mesosphere (e.g., Holton, 1983) as well as owing to interhemispheric stratosphere– mesosphere coupling as was demonstrated by model and experimental studies (Becker and Schmitz, 2003; Karlsson et al., 2007). According to these studies, it is possible that changes in the winter stratospheric zonal flow lead to altering of the mesospheric meridional circulation, which in turn (via adiabatic air expansion above the summer pole) results in temperature changes in the summer mesosphere. In particular, Shepherd et al. (2002) have demonstrated an inter-annual shift of about 5 days in a rapid enhancement of the upper mesosphere temperature during springtime in 1992 and 1993, that was supposedly connected to the last stratospheric warming event. However, we could not find in the literature a description of such pronounced inter-annual shift, on average, of 14 days in the summer mesopause temperature of 2007 which was, as we assume, due to a prominent event in the stratosphere– mesosphere dynamic coupling. This question will be addressed in further investigations. We have demonstrated that a general seasonal drop of the summer mesopause temperature determines the length and development of the NLC cycle. A detailed quantitative comparison between the NLC brightness (estimated from the ground) and Aura/MLS temperature behavior has been performed, unambiguously showing that the temperature decrease leads to increase of ¨ the NLC brightness. A similar result has been obtained by Lubken et al. (2009) based on a large statistics of model studies of mesospheric ice layers and background conditions from 1961 to 2008. In particular, the authors have demonstrated a rather strong anticorrelation of  0.69 between seasonal mean NLC brightness and temperature at 691N. Note that in the present study we obtained a moderate anticorrelation between nightly temperature and NLC brightness for some sites, namely, for Port Glasgow (  0.45) and Athabasca (  0.47) for 2007, and for Athabasca (  0.55), Novosibirsk (  0.51) and Vilnius (  0.54) for 2008. It has been shown that temperature variations explains the total variance of the NLC brightness by 5–22% for 2007 and 2–30% for 2008, depending on data set, and the average value of R2 equals to 14% and 15% for 2007 and 2008, respectively. This is 3–5 times more than the corresponding contribution due to temperature disturbances by the 2- and 5-day planetary waves. It is important to note that Aura makes almost instantaneous measurements of temperature profiles in the atmosphere while

2107

NLCs are usually observed over a 1–3 h period and therefore unregistered temporal temperature variations are readily possible. Also, the wind dynamics of the summer mesopause plays a significant role in assessment of the NLC brightness on a given night, that is, the NLC fields are sometimes seen to be brought from other regions of the mesopause (which is out of the field of view of a camera). Thus, even at an equally low temperature across a broad region of the mesopause, NLCs locally born in some region can be transported over a long distance and might be seen at other site, but it strongly depends on the speed and direction of wind which is highly variable in the summer mesopause both in time and space. We have studied a direct influence of the 5-day W1 and 2-day W3 planetary waves, extracted from the temperature fields by the Aura/MLS satellite data, on the NLC brightness variations observed from the ground. A 5-day period is well-known in NLC variability (Gadsden, 1985; Kirkwood and Stebel, 2003; Dalin et al., 2008) and is usually associated with a free traveling quasi 5-day planetary wave of zonal wavenumber 1. Note that theoretical considerations (e.g. Salby, 1981) demonstrate that the periods of 5-day planetary waves lie in the interval of 4.4–5.7 days (due to local nonuniformities and a Doppler-shift effect). This is confirmed by observational data, namely, by a 4–6-day period wind fluctuation in the upper mesosphere and mesopause (Jacobi et al., 1998) and by time intervals of 4–6 days between NLC displays around the globe (Dalin et al., 2008). Merkel et al. (2009) have employed the AIM (CIPS instrument) satellite and SABER instrument data to confirm a pronounced impact of the 5-day W1 and 2-day (W2 and E1) planetary waves on the PMC albedo and frequency for 2007. The maximum temperature amplitudes by the 5- and 2-day waves have been found as 3 K and 2.2–2.4 K, respectively. The maximum temperature disturbances obtained in the present study using the Aura/MLS data is 3 and 2.5 K for the 5-day W1 and 2-day W3 wave, respectively, that is the same values identified by Merkel et al. (2008, 2009). In this light, it is interesting to regard the study by Sandford et al. (2008) who have analyzed the Aura/MLS geopotential height data at a height of about 90 km for the period of 2004–2007 and have revealed significant activity of the 2-day planetary wave with zonal wavenumbers of W2, W3 and W4 during summertime in the northern hemisphere. Although these waves are most prominent at mid-latitudes, they are still significantly present at a 601N latitude circle. It is important to note that the authors have found the 2-day W1, 0, E1, E3 and E4 waves being very small in amplitude, and have not considered the activity of these oscillations in detail. Also note from this work that evolution of the 2-day wave of any regarded zonal wavenumber has essential seasonal and inter-annual variations. Thus, it is questionable so far what kind of the 2-day wave (what zonal wavenumber) dominates in the polar middle and upper atmosphere during summertime in the northern hemisphere; it seems that it depends on the data source and retrieving technique as well as on a particular season and year considered. As far as concerned NLC observations, it has been shown in Dalin et al. (2008) that the absolute number of the 2-day time interval between NLC displays was more than the absolute number of 4–6-day intervals in both 2006 and 2007. The larger number of the 2-day intervals points to the dominating activity of the 2-day planetary wave in the summer mesopause in 2006 and 2007, at least in the latitude band of 58–631N. Summarizing, it is clear that the 2-day wave (without regard to zonal wavenumber) plays an equally important role as the 5-day wave does, when considering NLC displays from the ground and PMC observations from space. The direct comparison of the 2- and 5-day wave patterns and NLC brightness variations demonstrates that there exists a

2108

P. Dalin et al. / Journal of Atmospheric and Solar-Terrestrial Physics 73 (2011) 2097–2109

negative trend for each observational site when temperature disturbances are plotted versus NLC brightness. For some individual sites the trend is 95% statistically significant, while for other sites it is not. But the combined data sets show up significant negative trend for both the 2- and 5-day wave analysis, and it is valid for both the 2007 and 2008 summer season. It is quantitatively expressed as follows. For 2007, the slope is equal to 0.1770.04 K/BU (standard deviation is 0.04) and  0.147 0.04 K/BU for the 2- and 5-day waves, respectively. For 2008, the slope is  0.1170.04 K/BU for both the 2- and 5-day waves; that is the 2- and 5-day wave produced approximately the same effect on the NLC brightness variations both in 2007 and 2008. It is necessary to note that Merkel et al. (2009) have demonstrated and claimed the 5-day wave is the dominant planetary wave in the summer mesosphere at high latitudes, and it has a noticeable effect on PMC albedo and frequency. The temperature amplitudes are up to 3 K due to this wave. At the same time, a close look at their results shows that the variability in PMC brightness and occurrence frequency explained by the 5-day wave is, in the best case, 10–12% and 22%, respectively. The effect of the 2-day W2 and E1 planetary waves is even less (7–9% and 12–16%, respectively). Our results demonstrate the variance of NLC brightness due to the 5- and 2-day planetary waves varies between 1% and 18% and between 1% and 13%, respectively, depending on longitude, with the average values of 3–4% due to the 5-day wave and 4–5% due to the 2-day wave. The average values of the NLC brightness variance seem to be by a factor of 2–3 smaller than the PMC albedo variance and it may be due to the following reason. The NLC brightness is visually estimated as the difference between the brightness of NLCs and of the background atmosphere. The PMC albedo is estimated as the ratio of atmospheric radiance to solar irradiance. So these are different scales. On the other hand, the NLC brightness is a logarithmic function of the light intensity due to the empiric psychophysiological Weber–Fechner law (Fechner, 1860). In this law there is a multiplicative constant (determined experimentally) which is unknown value in our case. Therefore, NLC brightness is not directly comparable to PMC albedo values. That is why the average values of the NLC brightness variance differ from the PMC albedo variance and probably due to this reason the range of the NLC brightness variance only partly belongs to the range of the PMC albedo variance found by Merkel et al. (2009). Thus, even if the 5- and 2-day waves are the dominant planetary waves in the summer mid-latitude and polar mesosphere, at the best these waves have just a moderate effect on the total variability of the NLC/PMC brightness and occurrence frequency. Other dynamical processes should explain the larger part of the NLC seasonal and global variability. The most probable candidates are small (wavelength of 5–10 km), medium (10–100 km) and large scale (100–1000 km) atmospheric gravity waves which produce larger temperature variations up to 5–10 K and even more (Rapp et al., 2002) around the summer mesopause. Indeed, we always observe gravity waves of small and medium scales, and often of large scales when NLCs appear. These waves highly modulate the brightness and morphology of noctilucent clouds during the period of visibility.

6. Conclusions By comparing the NLC observations obtained with the global ground-based network of NLC cameras and the Aura (MLS instrument) temperature and water vapor data, we have investigated NLC climatology and the impact of the 2-day westward (W3) and 5-day westward (W1) planetary waves on the NLC brightness variation during the summertime of 2007 and 2008.

A remarkable feature has been observed in the climatology of the summer mesopause at 601N in 2007, namely, the minimum of a Gaussian fitted seasonal temperature variation occurred on average by 14 days earlier than that in 2005, 2006 and 2008, as verified from zonal mean temperature behavior in 2005–2008. Also, the temperature curve minimum of 2007 has been broader than the corresponding one of 2008 by 5–28 days. It has resulted in an inter-annual shift of 12–29 days (depending on an observational site) of the maximum of a Gaussian fitted seasonal variation of the NLC brightness cycle occurred earlier in 2007. These prominent differences might be explained by the interhemispheric stratosphere–mesosphere dynamic coupling. We have shown that the MLS temperature values around the mesopause control the climatology of NLCs, that is the drop in the temperature below the frost point temperature in June corresponds to the beginning of the NLC season and rise of the temperature above the frost point temperature in August determines the end of the NLC season. The mesopause frost point temperature has been estimated to vary between 145 and 147 K (given variation of water vapor between 3.5 and 6.0 ppmv), and the statistical value of the temperature minimum has been found in the range of 138–140 K both for 2007 and 2008. In this sense, the MLS temperature can be used as an indicator for the duration of the NLC season. A detailed quantitative comparison, performed for the first time, between the NLC brightness and Aura/MLS temperature behavior demonstrates that decrease in temperature leads to increase in the NLC brightness. The temperature variations explain the total variance of the NLC brightness by, on average, 14% and 15% for 2007 and 2008, respectively. We have demonstrated that the 2- and 5-day planetary wave activity and NLC brightness variations tend to be in anticorrelation. When temperature disturbances are plotted versus NLC brightness, the regression analysis, applied for zonally averaged combined data of 2007, shows clear significant negative trend for both the 2- and 5-day wave, with the slope equal to 0.1770.04 and  0.147 0.04 K/BU, respectively. For 2008, the regression analysis applied for the combined data reveals clear significant negative trend as well, with the slope of  0.1170.04 K/BU for both the 2- and 5-day wave; thus, the 2- and 5-day planetary wave had nearly equal influence on the 2007 and 2008 NLC brightness variability. And finally, it is worth noting that the 2- and 5-day wave contribution into the total NLC/PMC variability is weak or moderate, on average 3–5%. Other factors, responsible for larger part of the day-to-day variability, should be sought, probably in the direction pointing to gravity waves of small-, medium- and large-scales; the first two categories are always and the last one rather often observed in the NLC field.

Acknowledgements We gratefully acknowledge Andrei Reshetnikov and Alexander Dalin for their support of the NLC observations with a digital camera in Moscow. The paper benefited from constructive comments and suggestions made by two anonymous reviewers and Guest Editor Jacek Stegman. The Aura/MLS data version 2.2 were obtained from the NASA Goddard Space Flight Center Data and Information Services Center (GSFC-DISK): http://disk.sci.gsfc.nasa. gov/Aura/data-holdings/MLS/ml2t.002.shtml.

References Becker, E., Schmitz, G., 2003. Climatological effects of orography and land–sea heating contrasts on the gravity wave-driven circulation of the mesosphere. Journal of the Atmospheric Sciences 60, 103–118.

P. Dalin et al. / Journal of Atmospheric and Solar-Terrestrial Physics 73 (2011) 2097–2109

Dalin, P., Kirkwood, S., Andersen, H., Hansen, O., Pertsev, N., Romejko, V., 2006. Comparison of long-term Moscow and Danish NLC observations: statistical results. Annales Geophysicae 24, 2841–2849. Dalin, P., Pertsev, N., Zadorozhny, A., Connors, M., Schofield, I., Shelton, I., Zalcik, M., McEwan, T., McEachran, I., Frandsen, S., Hansen, O., Andersen, H., Sukhodoev, V., Perminov, V., Romejko, V., 2008. Ground-based observations of noctilucent clouds with a northern hemisphere network of automatic digital cameras. Journal of Atmospheric and Solar-Terrestrial Physics 70, 1460–1472. Donahue, T.M., Guenther, B., Blamont, J.E., 1972. Noctilucent clouds in daytime: circumpolar particulate layers near the summer mesopause. Journal of the Atmospheric Sciences 29, 1205–1209. ˇ Dubietis, A., Dalin, P., Balcˇiunas, R., Cernis, K., 2010. Observations of noctilucent clouds from Lithuania. Journal of Atmospheric and Solar–Terrestrial Physics 72 (14–15), 1090–1099. Eckermann, S.D., Hoppel, K.W., Coy, L., McCormack, J.P., Siskind, D.E., Nielsen, K., Kochenash, A., Stevens, M.H., Englert, Ch.R., Singer, W., Hervig, M., 2009. Highaltitude data assimilation system experiments for the northern summer mesosphere season of 2007. Journal of Atmospheric and Solar–Terrestrial Physics 71, 531–551. Espy, P.J., Witt, G., 1996. Observation of a quasi 16-day oscillation in the polar summer mesospheric temperature. Geophysical Research Letters 23 (10), 1071–1074. Fechner, G.T., 1860. Elemente der psychophysik, 2 volumes. Druck und Verlag von ¨ Breitkopf und Hartel, Leipzig. Fogle, B., Haurwitz, B., 1966. Noctilucent clouds. Space Science Reviews 6, 279–340. Froidevaux, L., Livesey, N.J., Read, W.G., Jiang, Y.B., Jime´nez, C.C., Filipiak, M.J., Schwartz, M.J., Santee, M.L., Pumphrey, H.C., Jiang, J.H., Wu, D.L., Manney, G.L., Drouin, B.J., Waters, J.W., Fetzer, E.J., Bernath, P.F., Boone, C.D., Walker, K.A., Jucks, K.W., Toon, G.C., Margitan, J.J., Sen, B., Webster, C.R., Christensen, L.E., Elkins, J.W., Atlas, E., Lueb, R.A., Hendershot, R., 2006. Early validation analyses of atmospheric profiles from EOS MLS on the Aura satellite. IEEE Transactions on Geoscience and Remote Sensing 44 (5), 1106–1121. Gadsden, M., 1985. Observations of noctilucent clouds from North-West Europe. Annales Geophysicae 3 (1), 119–126. ¨ Gadsden, M., Schroder, W., 1989. Noctilucent Clouds. Springer, New York 165 pp. Holton, J.R., 1983. The influence of gravity wave breaking on the general circulation of the middle atmosphere. Journal of the Atmospheric Sciences 40, 2497–2507. ¨ Jacobi, Ch., Schminder, R., Kurschner, D., 1998. Planetary wave activity obtained from long-period (2–18 days) variations of mesopause region winds over Central Europe (521N, 151E). Journal of Atmospheric and Solar–Terrestrial Physics 60 (1), 81–93. ¨ Karlsson, B., Kornich, H., Gumbel, J., 2007. Evidence for interhemispheric stratosphere–mesosphere coupling derived from noctilucent cloud properties. Geophysical Research Letters 34, L16806. Kirkwood, S., Stebel, K., 2003. Influence of planetary waves on noctilucent clouds occurrence over NW Europe. Journal of Geophysical Research 108, 8440. Lambert, A., Read, W.G., Livesey, N.J., Santee, M.L., Manney, G.L., Froidevaux, L., Wu, D.L., Schwartz, M.J., Pumphrey, H.C., Jimenez, C., Nedoluha, G.E., Cofield, R.E., Cuddy, D.T., Daffer, W.H., Drouin, B.J., Fuller, R.A., Jarnot, R.F., Knosp, B.W., Pickett, H.M., Perun, V.S., Snyder, W.V., Stek, P.C., Thurstans, R.P., Wagner, P.A., Waters, J.W., Jucks, K.W., Toon, G.C., Stachnik, R.A., Bernath, P.F., Boone, C.D., Walker, K.A., Urban, J., Murtagh, D., Elkins, J.W., Atlas, E., 2007. Validation of the Aura Microwave Limb Sounder middle atmosphere water vapor and nitrous oxide measurements. Journal of Geophysical Research 112, D24S36. ¨ Lubken, F.-J., 1999. Thermal structure of the Arctic summer mesosphere. Journal of Geophysical Research 104 (D8), 9135–9149. ¨ Lubken, F.-J., Berger, U., Baumgarten, G., 2009. Stratospheric and solar cycle effects on long-term variability of mesospheric ice clouds. Journal of Geophysical Research 114, D00I06. Merkel, A.W., Thomas, G.E., Palo, S.E., Bailey, S.M., 2003. Observations of the 5-day planetary wave in PMC measurements from the Student Nitric Oxide Explorer Satellite. Geophysical Research Letters 30 (4), 1196. Merkel, A.W., Garcia, R.R., Bailey, S.M., Russell III, J.M., 2008. Observational studies of planetary waves in PMCs and mesospheric temperature measured by SNOE and SABER. Journal of Geophysical Research 113, D14202.

2109

Merkel, A.W., Rusch, D.W., Palo, S.E., Russell III, J.M., Bailey, S.M., 2009. Mesospheric planetary wave activity inferred from AIM-CIPS and TIMED-SABER for the northern summer 2007 PMC season. Journal of Atmospheric and Solar– Terrestrial Physics, 10.1016/j.jasp.2006.05.01. Pogoreltsev, A., 1999. Simulation of planetary waves and their influence on the zonally averaged circulation in the middle atmosphere. Earth Planets and Space 51, 773–784. ¨ ¨ Rapp, M., Lubken, F.-J., Mullemann, A., 2002. Small scale temperature variations in the vicinity of NLC: experimental and model results. Journal of Geophysical Research 107, 4392. Read, W.G., Lambert, A., Bacmeister, J., Cofield, R.E., Christensen, L.E., Cuddy, D.T., Daffer, W.H., Drouin, B.J., Fetzer, E., Froidevaux, L., Fuller, R., Herman, R., Jarnot, R.F., Jiang, J.H., Jiang, Y.B., Kelly, K., Knosp, B.W., Kovalenko, L.J., Livesey, N.J., Liu, H.-C., Manney, G.L., Pickett, H.M., Pumphrey, H.C., Rosenlof, K.H., Sabounchi, X., Santee, M.L., Schwartz, M.J., Snyder, W.V., Stek, P.C., Su, H., ¨ Takacs, L.L., Thurstans, R.P., Vomel, H., Wagner, P.A., Waters, J.W., Webster, C.R., Weinstock, E.M., Wu, D.L., 2007. Aura Microwave Limb Sounder upper tropospheric and lower stratospheric H2O and relative humidity with respect to ice validation. Journal of Geophysical Research 112, D24S35. Riggin, D.M., Liu, H.-L., Lieberman, R.S., Roble, R.G., Russell III, J.M., Mertens, C.J., Mlynczak, M.G., Pancheva, D., Franke, S.J., Murayama, Y., Manson, A.H., Meek, C.E., Vincent, R.A., 2006. Observations of the 5-day wave in the mesosphere and lower thermosphere. Journal of Atmospheric and Solar– Terrestrial Physics 68, 323–339. Rodgers, C., 1976. Retrieval of atmospheric temperature and composition from remote measurements of thermal radiation. Reviewer of Geophysical Space Physics 14, 609–615. Romejko, V.A., Dalin, P.A., Pertsev, N.N., 2003. Forty years of noctilucent cloud observations near Moscow: database and simple statistics. Journal of Geophysical Research 108 (D8), 8443. Salby, M.L., 1981. Rossby normal modes in nonuniform background configurations. Part II: equinox and solstice conditions. Journal of the Atmospheric Sciences 38, 1827–1840. Salby, M.L., 1982. Sampling theory for asynoptic satellite observations. Part II: fast Fourier synoptic mapping. Journal of the Atmospheric Sciences 39, 2601–2614. Salby, M.L., 1984. Survey of planetary-scale traveling waves: the state of theory and observations. Reviews of Geophysics and Space Physics 22 (2), 209–236. Sandford, D.J., Schwartz, M.J., Mitchell, N.J., 2008. The wintertime two-day wave in the polar stratosphere, mesosphere and lower thermosphere. Journal of Atmospheric Chemistry and Physics 8, 749–755. Schwartz, M.J., Lambert, A., Manney, G.L., Read, W.G., Livesey, N.J., Froidevaux, L., Ao, C.O., Bernath, P.F., Boone, C.D., Cofield, R.E., Daffer, W.H., Drouin, B.J., ¨ Fetzer, E.J., Fuller, R.A., Jarnot, R.F., Jiang, J.H., Jiang, Y.B., Knosp, B.W., Kruger, K., Li, J.-L.F., Mlynczak, M.G., Pawson, S., Russell III, J.M., Santee, M.L., Snyder, W.V., Stek, P.C., Thurstans, R.P., Tompkins, A.M., Wagner, P.A., Walker, K.A., Waters, J.W., Wu, D.L., 2008. Validation of the Aura Microwave Limb Sounder temperature and geopotential height measurements. Journal of Geophysical Research 113, D15S11. Shepherd, M.G., Espy, P.J., She, C.Y., Hocking, W., Keckhut, P., Gavrilyeva, G., Shepherd, G.G., Naujokat, B., 2002. Springtime transition in upper mesospheric temperature in the northern hemisphere. Journal of Atmospheric and Solar– Terrestrial Physics 64, 1183–1199. von Savigny, C., Robert, C., Bovensmann, H., Burrows, J.P., Schwartz, M., 2007. Satellite observations of the quasi 5-day wave in noctilucent clouds and mesopause temperatures. Geophysical Research Letters 34, L24808. Waters, J.W., Froidevaux, L., Harwood, R.S., Jarnot, R.F., Pickett, H.M., Read, W.G., Siegel, P.H., Cofield, R.E., Filipiak, M.J., Flower, D.A., Holden, J.R., Lau, G.K., Livesey, N.J., Manney, G.L., Pumphrey, H.C., Santee, M.L., Wu, D.L., Cuddy, D.T., Lay, R.R., Loo, M.S., Perun, V.S., Schwartz, M.J., Stek, P.C., Thurstans, R.P., Boyles, M.A., Chandra, K.M., Chavez, M.C., Chen, G.-S., Chudasama, B.V., Dodge, R., Fuller, R.A., Girard, M.A., Jiang, J.H., Jiang, Y., Knosp, B.W., LaBelle, R.C., Lam, J.C., Lee, K.A., Miller, D., Oswald, J.E., Patel, N.C., Pukala, D.M., Quintero, O., Scaff, D.M., Van Snyder, W., Tope, M.C., Wagner, P.A., Walch, M.J., 2006. The Earth Observing System Microwave Limb Sounder (EOS MLS) on the Aura satellite. IEEE Transactions on Geoscience and Remote Sensing 44 (5), 1075–1092.