View source

* Project leaders: G.Thomas (USA), U.Berger (Germany)
* Project members: S. Bailey (USA),G. Baumgarten(Germany),M. DeLand (USA), J. Fiedler (Germany),B. Karlsson (Sweden), S. Kirkwood (Sweden), A. Klekociuk (Australia), F.-J. Lübken (Gemany), A. Merkel (USA), N. Pertsev (Russia), J. Russell III (USA), E. Shettle (USA)
==Introduction==
The 20-yr old speculation that high-altitude summertime ice clouds (polar mesospheric clouds or noctilucent clouds, here denoted MC) are affected by  anthropogenic  activities has recently received support from a 30-year time series of SBUV (Solar Backscatter Ultraviolet) satellite measurements (see Figure 1).  SBUV data reveal a significant trend in bright MC properties. However, the robustness of the trend, extracted from interannual, local-time and solar-cycle variability, and its underlying causes remains debatable. It is important to understand the relative roles of these three factors  (solar, inter-annual and long-term forcing) before a definitive long-term trend can be evaluated. Furthermore the problem of attribution, that is, the nature of the various forcings on ice formation is not yet understood. For example with respect to the long-term changes, is a lowered temperature due to higher carbon dioxide responsible for the observed increase in brightness and occurrence frequency of MC? Or are water vapor changes due to oxidation of methane responsible, since we know that lower atmospheric methane has more than doubled over the past 120 years, and methane oxidation leads to upper atmospheric water vapor. The failure to detect any changes in the altitude of MC since the first measurement made by Otto Jesse in the late nineteenth century has provided an important constraint, since water vapor changes and temperature changes affect cloud altitude in different ways. Fortunately, the state of the art in modeling has now reached a point where ice formation is coupled with general circulation models. In a recent study Berger and Lübken (2011) showed that in the summer period 1979 -1997 at mid-latitudes strong cooling of up to 3-4 K/decade occurs in the middle mesosphere, in the period 1961-1979 the middle atmosphere cooled  significantly less, and for the period 1997-2009 they find a warming of ~1 K/decade.  For the first time, modeled temperature trends confirm the extraordinarily large temperature trends observed at mid-latitudes during the period 1979-1997 derived from lidar measurements, satellite data, and phase height measurements. The differences in temperature trends in the mesosphere originate from the evolution of stratospheric ozone in the past 50 years, e.g. the observed reversal of both stratospheric and mesospheric temperature trends in the mid 1990s is caused by the recovery of 
stratospheric ozone (WMO report 2011).  Therefore a new research question arises: does any trend in MC show a similar behavior? Relevant publications on this subject are reported in the references below:
[[File:Vqn-fig2rev.png|thumb|Fig. 1 A comparison of the seasonal PMC frequency of occurrence measured by SBUV and the fit to a linear regression in time and solar activity (upper panel) by latitude band and (lower panel) for all latitude bands combined between 54°N and 82°N. The error bars are the confidence limits in the individual seasonal mean values based on counting statistics, which do not reflect other factors such as inter-annual variability in large scale dynamics (from Reference 5)]]

==References==

#Luebken, F.-J., U. Berger, and G. Baumgarten (2009), Stratospheric and solar cycle effects on long-term variability of mesospheric ice clouds, J. Geophys. Res., 114, D00I06, doi:10.1029/2009JD012377 
#Merkel, A. W., Marsh, D. R., Gettelman, A., and Jensen, E. J.: On the relationship of polar mesospheric cloud ice water content, particle radius and mesospheric temperature and its use in multi-dimensional models, Atmos. Chem. Phys., 9, 8889-8901, 2009
# Merkel, A. W., D. Marsh, G. E. Thomas, C. Bardeen, M. Deland, WACCM simulations of long-term changes in polar mesospheric clouds, Layered Phenomenon in Mesospheric Regions (LPMR) Conference, Stockholm, 2009
#Marsh, D and A. W. Merkel, 30-year PMC variability modeled by WACCM, SA33B-08, Fall AGU Meeting, San Francisco, 2009
#Shettle,E. P.,M. T. DeLand, G. E. Thomas, and J. J. Olivero (2009), Long term variations in the frequency of polar mesospheric clouds in the Northern Hemisphere from SBUV, Geophys. Res. Lett., 36, L02803, doi:10.1029/2008GL036048.
#Thomas, G. E., D. Marsh and F.-J. Lübken,  Mesospheric ice clouds as indicators of upper atmosphere climate change, EOS, Transactions, American Geophysical Union, 91, No. 20, 18 May 2010, p. 183.
#Berger, U., and F.-J. Lübken (2011), Mesospheric temperature trends  at mid-latitudes in summer, Geophys. Res. Lett., 38, L22804, doi: 10.1029/2011GL049528.
#WMO (2011), Global ozone research and monitoring project-Report no. 52, Scientific assessment of ozone depletion: 2010, World Meteorological Organization


=Workshops and Meetings=
The first workshop under the auspices of this working group was held in Boulder, Colorado on December 10 & 11, 2010, entitled Modeling Trends in mesospheric clouds (Thomas et al, 2010).
[[file:COSTEP_LOGO_240.gif|left|]]
==[[Second CAWSES-2 Task 2  Workshop: Modeling Polar Mesospheric Cloud Trends, May 3-4, 2012]]==

Laboratory for Atmospheric and Space Physics, University of Colorado, USA.

We announce here a forthcoming workshop to be held in Boulder, CO USA on May 3-4, 2012. This is the 7th workshop sponsored by IAGA/ICMA/CAWSES  and focusses on global change in the upper mesosphere, specifically decadal-scale trends in Polar Mesospheric Clouds, PMC (or Noctilucent Clouds, NLC, as they are traditionally called when observed from the ground at twilight).

	The issue of long-term changes in PMC was addressed in the first workshop in Boulder, Colorado, 10-11 December, 2009 (Thomas et al, 2010). A long suite of satellite measurements by the Solar Backscatter Ultraviolet Spectrometer (SBUV) dating back to 1979 shows a significant long-term trend in mesospheric cloud activity and brightness.  Attribution of  trends in these high-altitude ice clouds (~83 km) includes:  (1) increasing carbon dioxide concentrations which tend to cool the upper mesosphere (the 80-90 km 'mesopause region'); (2) increasing water vapor due to growing concentrations of methane, in addition to changing efficiency of tropospheric water vapor entry through the tropopause; (3) shrinking of the upper atmosphere arising largely from decreased ozone levels and subsequent stratospheric cooling; (4) increased concentrations of water vapor due to the rise in space traffic over the last 30 years. Other possible causes are long-term changes in winds and wind filtering of gravity waves which dominate the dynamics of the high-latitude summertime mesopause region. An even larger (cyclic) trend is superimposed on the long-term record, believed to be a result of the solar cycle, but not well understood.

	General circulation models coupled with ice formation are now capable of simulating the SBUV trends to a remarkable degree, as discussed in the first workshop. However, new modeling results have brought out the importance of ozone trends, which affect the heat balance of the entire upper atmosphere. The LIMA/ICE model predicts a 'break-point' in the time series of PMC trends, due to the turn-around of the ozone trend in the mid-1990's. A wealth of new data from several satellite missions (AIM, ODIN, TIMED) are available. The Aeronomy of Ice in the Mesosphere (AIM) directly addresses the processes which control ice cloud evolution. In addition, more information is now available on diurnal variations in ice water content, which in principle affects the analysis of trends made from satellites in sun-synchronous orbits with slowly-varying local time coverage. These and other advances in modeling, data availability and trend analysis are motivations for a coming together of both modelers and data analysts to assess the current state of knowledge. A goal of the workshop would be to make recommendations for future work to resolve many of the outstanding issues.

	The PMC Trend Workshop will be held at the Laboratory for Atmospheric Physics at the Space Science Building, the newest LASP location at the University Research Campus Park, University of Colorado, Boulder, Co, USA on May 3 and 4, 2012. All interested researchers, and particularly students, are invited to the meeting. No registration fees will be charged. We have now committed the travel funds allocated to us by CAWSES.

	Gary Thomas (thomas@lasp.colorado.edu) and Uwe Berger (berger@iap-kborn.de), co-chairs of CAWSES-II, Task 2, Project 3

===Registration===

 '''The deadline for registration is closed.

===MEETING AGENDA (As of April 30, 2012)===

==AM Thursday, May 3== 

0830-0840:  	Gary Thomas, LASP
Welcome & Summary of Workshop Objectives

====[[Long-term Trends:  PMC Observations]]====
0840-0900
John Olivero, Embry-Riddle University, USA, Some Historical Notes on Noctilucent Cloud Studies

0900-0910 	Q&A, and discussion

0910-0930 	Matthew DeLand, SSAI, USA,Current PMC Trends Derived from SBUV Measurements

0930-0940 	Q&A, and discussion

0940-1000 	Mark Zalcik, Coordinator, NLC Can Am Network, Canada,Two Decades of Noctilucent Cloud Monitoring in North America

1000-1010 	Q&A, and discussion

1010-1025	P. Dalin, Swedish Institute of  Space Physics, Sweden,On the long-term trends in noctilucent clouds as observed from the ground and on the trends in the OH summer temperature as measured in Moscow and Lithuania

1025-1030	Q&A, and discussion

1030-1100 	Coffee Break

1100-1120 	Gerd Baumgarten, IUP, Germany,Decadal observations of particle sizes and water vapor content of NLC

1120-1130 	Q&A, and discussion

====[[Long-term Trends:  Temperature & Water Vapor Observations]]====

1130-1150 	Alain Hauchecorne, Laboratoire ATmosphères, France, Temperature trends in the stratosphere and in the mesosphere as seen from Rayleigh lidar observations

1150-1200 	Q&A, and discussion

1200-1220 	Karen Rosenlof, NOAA, USA , A new satellite based zonally averaged time series of stratospheric water vapor

1220-1230 	Q&A, and discussion

1230-1330 	Lunch Break
 
==PM Thursday, May 3==

====[[ Long-term Trends:  Temperature & Water Vapor Observations (cont.)]]====

1330-1350 	Gerald Nedoluha, NRL, USA , Long-Term Ground-based Microwave Measurements of Middle Atmospheric Water Vapor from NDACC sites

1350-1400 	Q&A, and discussion

1400-1420 	Michael Stevens, NRL, USA ,The impact of space shuttle main engine exhaust on PMCs and implications to trend studies

1420-1430 	Q&A, and discussion

====[[Long-term Trends:  Modeling]]====

1430-1450	Uwe Berger, IUP, Germany,Solar variability and trend effects in mesospheric ice layers

1450-1500	Q&A, and discussion

1500-1520	Aimee Merkel, LASP, USA,WACCM-PMC simulations of long-term trends of PMC

1520-1530	Q&A, and discussion

1530-1600	Coffee Break

1600-1620	David Siskind, NRL, USA,The PMC region as an integrator of coupling processes: Implications for trend studies from AIM and other missions

1620-1630	Q&A, and discussion

1630-1650	Stan Solomon,  NCAR, USA,Thermospheric Temperature Trends: Modeling and Observations

1650-1700	Q&A and discussion

1700-1720	Dan Marsh, NCAR, USA,Climate change in the mesosphere from 1850 to 2100 in CESM-WACCM

1720-1730	Q&A, and discussion

1730-1750	Kota Okamoto, The University of Tokyo, Japan, On the dynamical responses in the middle atmosphere to ozone recovery and CO2 increase

1750-1800	Q&A, and discussion

1800 	        Adjourn

1900-2100	Group Dinner (TBD)

==AM Friday, Friday, May 4==

====[[Inter-annual, hemispheric and seasonal variability:  Observations]]====

0830-0850	James Russell III, Hampton Univ, USA, AIM science results and their significance for PMC long-term change studies

0850-0900	Q&A, and discussion 

0900-0920	Cora Randall, LASP, USA, AIM/CIPS Observations of PMC Variability

0920-0930	Q&A, and discussion

0930-0945	Rachel Ward, Utah State University, USA, Comparison of Northern and Southern Hemisphere Mesospheric Gravity Waves using CIPS PMC Data

0945-0950	Q&A, and discussion

0950-1005	Susanne Benze, LASP, USA, On the onset of polar mesospheric cloud seasons

1005-1010	Q&A, and discussion

1010-1025	Hanli Liu, NCAR, USA ,Fast meridional transport in the lower thermosphere by planetary-scale waves

1025-1030	Q&A, and discussion

1030-1100	Coffee Break

====[[Properties of PMC and their Environment: Observations]]====

1100-1115	E.J. Llewellyn, University of Saskatchewan, Canada, Special Observational Opportunities Offered by PMCs - Nadir Observations in the Limb

1115-1120	Q&A, and discussion

1120-1135	Scott Robertson, U of Colorado, USA, Detection of Meteoric Dust in Mesosphere by the CHAMPS Rockets

1135-1140	Q&A, and discussion

1140-1200	Kristell Pérot, LATMOS Laboratoire ATmosphères, France, PMC Particle Size Retrieval from GOMOS / ENVISAT Observations

1200 -1210	Q&A and discussion

1210-1310	Lunch

==PM Friday, Friday, May 4==

====[[Properties of PMC and their Environment: Modeling]]====

1310-1330	Charles Bardeen, NCAR, USA, Simulations of PMC during the AIM time period using WACCM/CARMA

1330-1340	Q&A, and discussion

1340-1400	Mark Hervig, GATS, Inc., Variability in PMCs and their environment from SOFIE observations, and potential implications for PMC trends

1400-1410	Q&A, and discussion

1410-1425      I. Azeem, Astraspace, USA, Simulations of Shuttle Main Engine Plume Effects on Lower Thermosphere Energetics and Chemistry

1425-1430	Q&A, and discussion

1430-1445	Coffee Break

Panel and group discussion

1445-1545	Panel discussion

1545-1630	Concluding remarks & summary from the project co-chairs
	
1630   Adjourn

==Local Organizing Committee: Aimee.Merkel@lasp.colorado.edu,Dan Marsh (marsh@ucar.edu) and Charles Bardeen (bardeenc@ucar.edu)==

===NOTE: The Particle Size Workshop,originally scheduled for May 2, has been postponed to a later date.===

==SPECIAL JOURNAL ISSUE ON UPPER ATMOSPHERIC TRENDS NOW AVAILABLE. The joint JGR/Space, JGR/Atmospheres special section on upper atmospheric trends is now complete, and can be accessed at http://www.agu.org/journals/ja/special_sections.shtml?collectionCode=UATREND1&amp;journalCode=JA==

='''Observing Facilities'''=


== Aeronomy of Ice in the Mesosphere (AIM), a NASA satellite mission (2007-) ==
[[File:AIMforWIKI.png|left|Fig. 2 Artist's conception of the AIM spacecraft in orbit, showing the line of sight of the SOFIE solar occultation experiment (courtesy, J. Russell III]][[File:CIPSfig1.png|right|]]AIM was launched from Vandenberg Air Force Base on April 25, 2007 becoming the first satellite mission dedicated to the study of Polar Mesospheric Clouds (PMCs).  A Pegasus XL rocket placed the AIM satellite into a near circular (601 km apogee, 595 km perigee), 12:00 AM/PM sun-synchronous orbit. By measuring PMCs and the thermal, chemical and dynamical environment in which they form, AIM will quantify the connection between these clouds and the meteorology of the polar mesosphere. In the end, this will provide the basis for study of long-term variability in mesospheric climate and its relationship to global change. The results of AIM will be a rigorous test and validation of predictive models that then can reliably use past PMC changes and current data to assess trends as indicators of global change. This goal is being achieved by measuring PMC densities, spatial distribution, particle size distributions, gravity wave activity, meteoric smoke influx to the atmosphere and vertical profiles of temperature, H2O, O3, CH4, NO, and CO2. 

The overall goal of AIM is to resolve why PMCs form and why they vary. It has been suggested that the observed changes in the clouds are related to increased concentrations of greenhouse gases. This suggestion is plausible because an increase in carbon dioxide, while warming the surface of the Earth, cools the upper atmosphere which can facilitate mesospheric cloud formation there. Additionally, increases in methane at the surface of the Earth lead to increases in water vapor at high altitudes through chemical oxidation processes, which further facilitates cloud formation and growth. While plausible, this greenhouse gas hypothesis has not yet been proven.The AIM webpage is at [http://aim.hamptonu.edu/]


----

== Arctic Lidar Observatory for Middle Atmosphere Research (ALOMAR) ==
[[File:alomar-rmr-lidar-small.png|left|Fig. 1:ALOMAR observatory and laser beams of the ALOMAR RMR-lidar during operation with tilted telescopes (courtesy, J. Fiedler)]] [[File:rmr-timeseries-small.png|right|]]Fig. 2:Year-to-year variability of seasonal mean NLC occurrence and altitude for two different cloud classes. The blue curves contain all measurements having a sensitivity above the long-term detection limit, whereas the green curves show results for strong clouds only. The vertical bars indicate 95% confidence limits for the occurrence and errors of the mean altitudes. For more information see text and references.
The Rayleigh/Mie/Raman (RMR) lidar at the Arctic Lidar Observatory for Middle Atmosphere Research (ALOMAR) is located on top of a ≈400 m high mountain in Northern Norway (69.3°N, 16.0°E). It was installed in 1994 and designed for multi-parameter investigations of the Arctic middle atmosphere on a climatological basis. Because of its Arctic location, the lidar is optimized for measurements during full sunlight. Appropriate technical solutions, laser wavelength stabilization in combination with strong spectral as well as spatial filtering at the receiving system, have been implemented. The lidar is a complex twin-system consisting of two power lasers, two receiving telescopes, and one optical bench for spectral separation and filtering of the light received from the atmosphere. The lasers emit pulses at three wavelengths (355 nm, 532 nm, 1064 nm) simultaneously with an overall peak power of 150 MW per laser. The backscattered light from the atmosphere is collected by telescopes with a diameter of 1.8 m each. They can be tilted up to 30° off-zenith to allow different viewing directions (Fig. 1), which separates the sounding volume up to 90 km at an altitude of 80 km. After spectral and intensity separation the light is analyzed by 15 channels and recorded by single photon counting detectors. 
One main objective of the RMR-lidar is the observation of ice layers in the mesopause region, which are known as noctilucent clouds (NLC) and polar mesospheric clouds (PMC). Throughout the NLC season (1 June to 15 August) the lidar is operational for 24 hours per day to measure whenever permitted by the local weather conditions. This yielded a total of more than 4000 measurement hours from 1997 to 2009. NLC were observed during ≈1700 hours which is the largest NLC data base acquired by lidar. The data are used to investigate decadal scale changes of NLC parameters, size and number density of the ice particles, as well as small scale structures which are often observed in the cloud layers. Fig. 2 shows the time series of NLC occurrence and altitude above ALOMAR covering one solar cycle (taken from reference 2). During these 11 years there is no statistical significant anti-correlation between cloud occurrence and solar activity, which is partly in contrast to other data sets. The mean cloud altitude is 83.2 km and appears to remain nearly unchanged since the first NLC observations more than 100 years ago.
----

==Visual Observations of Noctilucent Clouds ==



[[File:2005-06-24_213509-GB-2005-06-24_213551-GB-sm.png|center|Fig. 1:Noctilucent cloud display seen from Kühlungsborn, Germany, on June 24, 2005 while the sun is about 8 degrees below the horizon. (courtesy, G. Baumgarten)]] Scientific interest in noctilucent clouds (NLC) can be traced back to 1884 when many observers watched the twilight sky to see the dramatic sunsets caused by dust from Krakatoa, which had erupted during the previous year. Captivating displays of 'night shining' clouds,  were seen and quickly recognized as lying at much higher altitude than normal clouds.  For much of the 20th century, visual and photographic observations were the only methods available for systematic monitoring of NLC characteristics. By the early 1960's it was clear that their occurrence rates varied enormously from year to year. Widespread, intense displays of NLC were seen for a few  consecutive summers, with NLC reported from somewhere around the 50 - 60 N latitude band almost every night over the summer season. These periods were followed by years when almost no NLC at all were seen. The most notable 'high spots' for NLC were the years 1885-1890 and 1963-1968, which were each followed by strong declines in NLC reports even though the same observers as during the 'hot spots' continued to look for them (reference 1).

Starting with the International Geophysical Year in 1958,  professionally organized observing networks started to gather systematic records of visual sightings of  noctilucent clouds (see references). Even though professional involvement has generally ended, or become sporadic, systematic NLC observations are still  collected by networks of enthusiasts, in Europe, Russia, Canada and North America.  Since 1996, visual observations by members of the public are collected at a number of regional or national centers [http://www.kersland.plus.com/nlcrepor.htm#nlccanam], and since 1996 at the 'Noctilucent Cloud Observers Homepage' [ http://www.kersland.plus.com/].  NLC observations  from the UK and Denmark since 1964 form the longest continuous record (observing latitudes from 51 - 61 N). These show how NLC are most common in years of low solar activity, and rare in years of high solar activity. They do not show any significant increase over the last 45 years although a few percent increase (or decrease) cannot be ruled out (Fig. 2, reference 3).
[[File:NLC_UK_Denmark.png|200px|thumb|right|Fig. 2 A comparison of solar activity and the seasonal NLC frequency of occurrence according to reports of visual observations from the UK and Denmark (from Reference 3, extended to 2009 using internet reports [http://www.kersland.plus.com/]) ]]

Although monitoring of large-scale structures and possible trends in NLC is being taken over by satellite measurements, visual and photographic observations still have an important role to play. For example, they are the best method available for studying fine-scale structure, on scales of 10s of km or less, and they can be instrumental in identifying NLC at unusually low latitudes where satellites observations and ground-based remote sensing instruments are not available. 

[[File:2005-06-24_213509-GB-sm.jpg|left|Fig. 3:Wave structures on scales of several km are often seen in noctilucent cloud and highlight the dynamical processes leading to the cold summer mesopause. (courtesy, G. Baumgarten)]] 

#Fogle,B.and B.Haurwitz,Long term variations in  noctilucent cloud activity and their possible cause,  in Climatological Research,edited by K.Fraedrich,  M.Hantel,H.Claussen Korff,and E.Ruprecht,pp. 263–276,Heft 7,Bonner Meteorologische Abhandlungen.,Bonn,Germany,1974. 
#Romejko, V.A., P.A. Dalin, N.N. Pertsev, Forty years of Noctilucent Clouds observations near Moscow: database and simple statistics, J. Geophys. Res., 108, D8, 8443, doi: 10.1029/2002JD002364, 2003.
#Kirkwood, S., P. Dalin, A. Rechou, Noctilucent clouds observed from the UK and Denmark – trends and variations over 43 years, Annales Geophysicae, 26, 1243-1254, 2008.
----



== Additional notes about NLC long-term behavior according to visual observations ==


1. There is no distinct contradiction when comparing a trend in the NLC occurrence frequency from ground-based observations and a trend in the PMC satellite observations for lower latitudes (50-64°N), if we take the same years for an analysis. The difference between the NLC and PMC trend is rather in their significance probability. However, if we consider ground-based NLC observations for more than 40 years (which are longer than space-borne measurements) we arrive at almost zero trend in the NLC occurrence frequency and at very small positive trend in the NLC brightness which have no statistical significance.

[[File:Pertsevfig_1.png|200px|thumb|right|Fig. 1 Residuals (after subtracting the variation correlated with the solar Lyman alpha flux) of the normalized NLC frequency. The thick line represents the secular trend and its 95% confidence interval. The upper panel demonstrates the Moscow linear fit for 1962–2005, the central panel is for the Moscow linear fit for 1983–2005 and the lower panel is for the Danish data for 1983–2005 (courtesy N. N. Pertsev).]]
 
2. NLC most frequently occur in 1-2 years after the sunspot minimum and this delay is statistically significant. The PMC observations show a smaller delay of 0.5±0.5 year.

3. Concerning the year of the NLC discovery, we should note the following. The majority of papers, devoted to first observations of noctilucent clouds, refer not to 1884 but to 1885 as a year of first reliable descriptions undoubtedly concerning noctilucent clouds. Leslie (1884) did describe some sky phenomenon, resembling noctilucent clouds, but in his successive papers, Leslie (1885) and (1886), devoted to luminous clouds as a new phenomenon, he wrote nothing about priority of that observation of 1884; moreover, he did not refer to it at all. So, the description of Leslie's observation of 1884 cannot be regarded as reliable. Summarizing, Gadsden and Schröder (1989) wrote in their canonical book: “…It is certain that at the times of coloured twilight appearances of 1883/1884, no noctilucent clouds were discovered. Various reports also exist which could be interpreted as noctilucent clouds, but this will always remain uncertain (Pernter 1889; Schröder 1975; Gadsden 1985).”. 

References:

#Dalin, P., S. Kirkwood, H. Andersen, O. Hansen, N. Pertsev, V. Romejko, Comparison of long-term Moscow and Danish NLC observations: statistical results, Annales Geophysicae, 24, 2841-2849, 2006.
#Leslie, R., 1884. The sky-glows. Nature 30, 583.
#Leslie, R., 1885. Sky glows. Nature 32, 245.
#Leslie, R., 1886. Luminous clouds. Nature 34, 264.

P. Dalin, N. Pertsev