Reconstruction of height and position of the green airglow from July 15th, 2012

It was an ironic situation when during the night from 14^th to 15^th of July 2012 (at a weekend) a high number of observers and photographers were looking for a predicted aurora borealis and instead were confronted with a remarkable outburst of structured (or banded) green airglow. This phenomenon is well known and explored by professional geo-scientists but seemed to have slipped the attention of most amateur observers, including myself, up to then. Though it first seemed likely that the geomagnetic storm may have somehow triggered this event, later observations (e.g. July 23^rd, 2012: http://www.polarlichter.info/airglow.htm) indicated that the traditional excitation mechanism (UV and X-Ray radiation from the sun) is capable of producing intense green airglow without the need for a geomagnetic anomaly.

Due to the fascination I felt during my own observation, I got interested in using the many available photographs from the July 14^th/15^th night for a height and position reconstruction. However, as I later found out from literature, the airglow height of  87-95 km (i.e. a quite thin layer, comparable to the NLC layer around 83 km) is already well established by professional measurements. It is remarkable that this value can in fact be reproduced by comparing amateur photographs from various locations in Germany by an un-biased analysis, which I want to present here.

The first task to do was to contact other observes via the well-known communication boards about atmospheric optics to gather suitable photographic material. Of course I had my own images at hand and intended to use them for this process, so I already had a list of time slots to find synchronous counterparts for. Even though I could find several pictures taken within a tolerance of < 1 minute with respect to my own photos, I had to drop most of them since a coarse analysis of the viewing directions yielded no overlapping fields of view. But through discussing my idea with several other photographers, I was able to identify other matching pairs independent from my own material. Finally, I ended up with two data sets (image pairs), 1 and 2,  to work with:

1a) Frank and Sabine Wächter: July 15^th, 00:55 CEST, 51° 12’ N, 13° 35’ E, 189 m above sea level (Meißen, Saxonia): https://dl.dropbox.com/u/8849406/Forum/AirglowBlog/1a.jpg

1b) Jens Hackmann: July 15^th, 00:55 CEST, 49° 29’ N, 9° 55’ E, 333 m above sea level (Weikersheim, Baden-Württemberg): https://dl.dropbox.com/u/8849406/Forum/AirglowBlog/1b.jpg

2a) Franz Peter Pauzenberger: July 15^th, 02:02 CEST, 49° 00’ N, 11° 30’ E, 518 m above sea level (Beilngries, Bavaria): https://dl.dropbox.com/u/8849406/Forum/AirglowBlog/2a.jpg

2b) Alexander Haußmann: July 15^th, 02:01 CEST, 51° 32’ N, 13° 58’ E, 110 m above sea level (Senftenberg, Brandenburg): https://dl.dropbox.com/u/8849406/Forum/AirglowBlog/2b.jpg

For a detailed analysis, it is necessary to calibrate these photos, which means to precisely assign values for azimuth and elevation to each pixel. If the projection characteristics of the photographic lens are known, the positions of two stars in each picture are sufficient input for this purpose. However, the simple assumption of an ideal gnomonic (rectilinear) or equal-area projection (for ordinary and fisheye lenses, respectively) drastically limits the accuracy of the results. To overcome this, the projection characteristics for all four lenses were reconstructed by measuring the pixel distances of approximately 15 stars from the image center and compare these with the angular distance from the optical axis for each image.

After this calibration and assignment, longitude and latitude positions for each pixel can be calculated, allowing the projection of the photo onto a map if a certain height of the airglow layer is assumed. This method already proved to be very useful for the reconstruction of NLC positions (http://www.meteoros.de/php/viewtopic.php?t=8451). Since the goal is here to determine the layer height, this parameter is varied until the corresponding structures in both reconstructions of an image pair give the best fit. Indeed it was possible to find consistent height values for both data sets, 92 km for pair 1 (https://dl.dropbox.com/u/8849406/Forum/AirglowBlog/1.gif) and 93 km for pair 2 (https://dl.dropbox.com/u/8849406/Forum/AirglowBlog/2.gif). Here the traditional blink comparison technique was applied in a modern form using gif animations. It is fascinating to see how the airglow structures that look completely different in the original two photos of each pair coincide in the reconstruction on the map. Evidently, all non-airglow structures such as trees, background light, clouds, photographic violet aurora etc. have to be ignored in the reconstruction. It should be noted that more complex approaches (http://www.opticsinfobase.org/ao/abstract.cfm?uri=ao-51-7-963) are recently established in the professional field, allowing even to resolve finer structures within the thickness of the airglow layer.

Furthermore, these reconstructions show an undistorted view on the band structure of the green airglow layer. As already expected from the perspective view of the original photos, these bands are roughly aligned in the direction from West to East. Using the consistent height information obtained from the image pair comparisons, it is moreover justified to project a whole picture series from a single observation site onto the map in order gain insight in the airglow band dynamics. For this purpose I used a time lapse series that I took from July 14^th, 23.16 CEST to July 15^th,  01.20 MESZ at the Senftenberger See (51° 29’ N, 14° 01’ E), starting in the evening dawn and originally intended to capture the predicted aurora borealis (https://dl.dropbox.com/u/8849406/Forum/AirglowBlog/3.avi). Due to the weak contrast of the airglow at this stage, strong image processing is needed to separate the bands from the background. Though this finally results in a rather poor signal to noise ratio, it can clearly be seen that the airglow bands move in northward direction (https://dl.dropbox.com/u/8849406/Forum/AirglowBlog/4.avi), illustrating the recombination and/or matter transport dynamics in the mesopause region.

Author: Alexander Haussmann