The Aurora Alarm
Home FAQ Tech Data Gallery


Confirmed auroras for which alarms were successfully issued
Walla Walla, WAUrbana, IL
11April 1997(1) Not yet operational
4 May 1998(3)
16July 1998(1)
23July 1998(1)
22August 1998(1)
Out of order18February 1999(1)
10March 1999(2)
17April 1999(2)
15October 1999(1)
22October 1999(1)
7 April 2000(1)
24May 2000(3) 24May 2000(3)
Out of order12August 2000(1)
29October 2000(1)
28November 2000(3)
29November 2000(2)
20March 2001(2)
31March 2001(3) 31March 2001(3)
1 April 2001(1)
18April 2001(1)
18June 2001(3)
Out of order22October 2001(1)
6 November 2001(3)
24November 2001(1)
4 September2002(1)
8 September2002(1)
25October 2002(1)
2 November 2002(1)
21November 2002(2)
22January 2004(3*) Decommissioned
13October 2004(1)
8 November 2004(3)
9 November 2004(2)
10November 2004(2)
30May 2005(1)
12July 2005(1)
17July 2005(1)
9 August 2005(1)
Total: 34 Total: 9

(*) Alarms logged, but email not successfully sent

Recent News

4 February 2014: Service discontinued

Please see the announcement on the front page.

16 July 2007: Mailing lists moving again

Walla Walla College is changing its name to Walla Walla University, and its domain name is changing too. On Monday, July 16th, the mailing lists will be moved to No action is required on the part of subscribers, but the URLs for subscribing, unsubscribing, or changing membership preferences will now be and

2 November 2006: Alarm e-mails shortened

Cingular has been splitting the alarm messages, monthly tests included, into four tiny SMSes when delivering to my phone. That way they can charge me $0.40 for each event (test or real alarm). I got tired of this and changed the message body to a single line, either

The aurora is visible NOW.  See
I've updated the examples elsewhere on this site to reflect the change. The format of the subject line is not changing.

9 July 2006: Mailing lists moving

Effective today, the aurora-northwest and aurora-announce mailing lists have relocated to a server at Walla Walla College. The new URLs for the list info pages are and

All existing subscribers have been transferred to the new database, so all you should notice is the new sender address. The main project web site is not moving (for now).

22 January 2004: Northwest detector, mailing lists are back

If you've checked this web site lately, you may have noticed that the northwest aurora detector is back online and providing data. Everything is set up on a new domain,; you were redirected here if you pointed your browser at, the old site.

The changeover involved quite a few edits of various config files, which I made... except for the script that actually mails the alarm messages to the list (oops). And that's why you didn't receive email last night (Wednesday/Thursday) during the fantastic aurora. Sorry! You can still see the data, though:,2004-01-22

Thanks for your patience during this transition. I have settled for now in Pleasantville, New York, and am working on medical ultrasound projects at Philips Research. (Obligatory disclaimer: The opinions expressed on this site are my own, not my employer's.)

28 July 2003: Midwest detector going away; other services interrupted

The Urbana, IL aurora detector and the aurora-midwest mailing list service will be deactivated in late August. I finished my Ph.D. in May and am moving back to the west coast after seven years in Illinois. I did consider various alternatives for siting the detector and transferring responsibility to other people, but no acceptable solutions were found. Perhaps the hardware can be set up somewhere else when I settle down again.

Due to my uncertain employment situation, other services (like the aurora-northwest mailing list) may be interrupted for an indefinite period. The basic, static content of these web pages should remain accessible during this time.

Many aurora-midwest subscribers may wish to sign up for the aurora-northwest mailing list. I will not automatically transfer the subscriber list, so you will need to use the Mailman interface. Don't worry about unsubscribing from aurora-midwest; that entire list will go away after I remove the detector from the field site. I thank the Department of Electrical and Computer Engineering at the University of Illinois for providing space for the aurora alarm over the last four and a half years.

24 October 2002: Both detectors back online

I'm pleased to announce that both detector stations are back online. The Walla Walla detector, which has been operating but not sending data for more than a year, appears to be in good shape. You can even see a sub-threshold aurora in last night's data between 2 and 5 a.m. PDT.

11 January 2002: Northwest detector offline until spring

See my post to aurora-announce for details.

6 January 2002: Mailing list changes

I have switched all of the mailing lists over to new software (Mailman) that provides better functionality than Majordomo. Among other improvements, you can now manage your subscriptions on the web. The aurora-talk list has also been renamed aurora-announce, to better reflect its purpose. See my post to aurora-announce for details.

23 November 2001: Web page updates; northwest detector status

The gallery page has been updated with my photographs from the November 5-6 aurora. Despite the full moon, this display showed some beautiful, vivid colors.

The northwest aurora detector has been out of order for quite some time now, and we've missed warning on several good displays visible from that part of the country. The problem has been traced to a bad radio in the telemetry link, but a fix has been slow in coming due to the inaccesibility of the site and the fact that this is a spare-time effort. Hopefully it will be repaired before winter sets in.

9 August 2001: New web page, name change

After the major activity this spring I decided it was finally time for a web site re-design. The new layout is hopefully more attractive and easier to navigate. The name has also been changed from "The Aurora Monitor Project," partly to avoid confusion with the excellent, but unrelated, software package by the same name.

1 April 2001: Huge aurora visible across most of North America

My sincere thanks to those who have posted their aurora viewing reports to the list. Obviously this was a major event, with many confirmed sightings down to the Mexican border. I hope the aurora alarms aided many of you in your quest to see this beautiful show. A report of my observations from just north of Champaign-Urbana, Illinois (40.12 N 88.24 W), along with photographs, can be found on the gallery page.

30 October 2000: New mailing list: aurora-talk

In order to keep the alarm mailing lists quiet for those who subscribe using pagers and cell phones, a new mailing list has been created: aurora-talk. From now on, nothing but alarm messages and the monthly automatic test will be sent to aurora-midwest and aurora-northwest. Everyone is encouraged to subscribe to aurora-talk using their "normal" e-mail address.

5 October 2000: Northwest aurora detector back online

Finally, after extensive repairs, the aurora detector overlooking the Walla Walla valley in Washington is back online, thanks to the efforts of Dr. Rob Frohne, an engineering professor at Walla Walla College, and myself. You can see pictures of the new enclosure (a 30 gallon Rubbermaid garbage can) below. We hope this will last another few years :)

5 October 2000: Aurora monitor pages are Slashdotted!

Last night the Aurora Monitor web page was "slashdotted," meaning that it was featured in a brief news item at which is a widely read web site with about a million hits per day. You can see the story yourself here. After this story was posted, traffic on my web server increased thirty-fold and the total readership of the mailing lists is now 701 (up from 470 a week ago). The breakdown by region is 435 for aurora-midwest, 266 for aurora-northwest.

31 August 2000: Northwest aurora detector repaired (mostly)

This week I have been busy repackaging the aurora detector which overlooks Walla Walla, WA. The old enclosure had been entered by birds and wasps. Tunnels and "rooms" with straw, pine needles, etc. had replaced most of the original insulation. Some water was able to enter the electronics enclosure as well, causing minor corrosion which was apparently the cause of the outage last spring. All in all, we were quite satisfied that the old packaging managed to last this long in a mountain-top environment (about 4 years).

The good news is that everything has been resealed and insulated. The new enclosure (a 30 gallon garbage can) is filled with spray foam and foam packaging peanuts, with chicken wire and screen barriers to prevent unwanted guests from making their homes inside. I had hoped to make the trip up the mountain to re-install everything this week, but I ran out of time and have to leave for Oregon today. After this weekend, someone else will take it up the mountain and turn it on, and we hope to have data flowing again by sometime next week.

12 August 2000: Aurora makes appearance during Perseid meteor shower

It was timed perfectly to match the *only* dark hour (no moon) on the peak night of the Perseids meteor shower, so many of the amateur astronomers who went out for the meteor shower got to watch the aurora also. I viewed this one from my porch, since there was only another hour before twilight (see the gallery page). This was my sixth aurora sighting; now I'm yearning for one "real aurora" with some bright, rapidly moving forms.

16 July 2000: Huge aurora goes mostly unseen

Yesterday's magnetic storm may go down as the largest of solar cycle 23, but it started too early for North America. It seemed to be holding up through the afternoon, but as sunset approached there was a sharp decline in activity, and there was almost no aurora for North American viewers. I actually did see it for perhaps 5-10 minutes, out of 3+ hours of watching (see the gallery page). The Urbana aurora detector showed a very small blip, but nothing strong enough to trigger an alarm. I'd say it worked correctly; what I saw was hardly worth the drive out into the country.

7 June 2000: Northwest aurora detector down for the summer

After almost four years of near-flawless operation in a harsh mountaintop environment, the northwest aurora monitor appears to have failed last week. There were some problems with intermittent data over the last several days, and when a longer string of data was received last night, it showed the photomultiplier tube output at full scale all night long, causing several false alarms. It is unclear whether this is due to a bad tube or a problem somewhere else in the system.

This is good timing in some ways; the site is accessible during the summer, making repairs a possibility without having to wait through the winter. Also, as we approach the summer solstice there are hardly any hours of darkness left for aurora hunting, at least in Washington and British Columbia. Unfortunately, the repairs will have to wait until the last week of August when I visit the area. Thanks for your patience. I'll do my best to have things working again by the first of September.

24 May 2000: Aurora visible in central Illinois

A very nice aurora was seen last night from just north of Champaign, IL (and many many other places). From here it was green topped with red, with occasional rays extending above Polaris. See the gallery page for details.

22 October 1999: Bright aurora during full moon

Last night saw the strongest magnetic storming of the past year. There were K indices of 8 over two consecutive three-hour periods, from 8 p.m. to 2 a.m. PDT---ideally placed for North American viewers. The activity was strong enough to be clearly visible in the Urbana, Illinois data---the first major event to be seen from this location since data collection started in February.

Unfortunately, on any given night, there is always a chance that moonlight will spoil the show, and this was the case last night. The alarm thresholds that I use are biased strongly upwards by moonlight, making it very difficult to trigger an alarm when the moon phase is anywhere near full. The one alarm that was issued for the northwest did not occur until moonset, just before dawn. No alarm was generated for the midwest.

[The alarm thresholds have since been relaxed slightly to discriminate less against very bright moonlight --mh]

15 March 1999: More auroras

Solar cycle 23 continues! The aurora monitor near Walla Walla, WA has detected five auroral displays since the extensive aurora of 3-4 May 1998. These occurred on the evenings of 15-16 July, 22-23 July, 21-22 August (visible from the Oregon Star Party east of Bend), 24-25 September, and 9-10 March. Some of these were quite faint; an additional two displays on 26-27 August and 18-19 February were missed due to data outages. There has been a decided lull in solar activity over the past several months, so another surge is likely anytime now. We can hope that this will take place just in time for spring and better weather.

15 March 1999: Hardware status

A second aurora monitor is now operating near Urbana, IL. This unit is exactly the same as the first one still operating near Walla Walla, WA, except that a type 931B photomultiplier tube is being used (the Walla Walla instrument uses a 1P28 PMT). At this site we are fortunate to have a direct connection to the Internet, so an old PC running Linux is all that is required on-site, instead of the complicated radio telemetry system needed in Washington.

This addition has provided the chance to experiment with dial-up telephone notification for getting the alarm messages out. I am currently using a voice-capable modem with mgetty and the vgetty voice mail extensions running on my Linux machine at the University of Illinois. This system wakes me up at home if an alarm is issued. If anyone would like to experiment with this kind of set-up, let me know. If you live in the Champaign-Urbana local calling area, I may be able to add you to the telephone list.

The original aurora monitor overlooking Walla Walla, WA has now survived three winters without a major breakdown. Its built-in clock is now running far behind and it is likely that water has seeped into the outer enclosure, so a maintenance visit will probably be necessary either this summer or the next.

January 1999: Fall AGU

Last December I had the opportunity to travel to San Francisco for the fall meeting of the American Geophysical Union. The aurora monitor was featured in a poster presentation for a special session devoted to education and public outreach projects in magnetospheric physics. I had a wonderful time and the poster was well received. Some attendees expressed an interest in developing an inexpensive aurora detector suitable for distribution to school groups. The goal here would be to eliminate the costly photomultiplier tube and high voltage supply in favor of photodiodes and some form of lock-in detection. Future collaboration with other public outreach efforts like The INSPIRE Project is also possible.

7 May 1998: Big aurora!

On Sunday evening, May 3 (May 4 UT), the strongest magnetic storm to hit the earth in several years caused widespread appearances of the northern lights as far south as the northern and central United States. The aurora monitor was successful in sensing this display and generating automated warnings. See the gallery page for more details.

11 April 1997: Second aurora detected

The aurora monitor detected its second display of the northern lights last night! Since last November, data has been collected and analyzed more or less constantly, allowing the effects of cloud cover, precipitation, moonlight, and even the Geminids meteor shower to be quantified. By January I felt comfortable enough with the monitor's behavior to write a simple data-interpreting Perl script which runs on my computer at the University of Illinois. By placing the alarm functionality here, I have been able to tweak alarm thresholds, monitor the performance of the alarm algorithms, and note any false alarms on a day to day basis. Perhaps the best news, however, is that the aurora monitor has successfully endured a harsh winter in the Blue Mountains and is still sending data.

Last night (10-11 April) between 9 p.m. and 12:30 a.m. PDT, any doubt about the effectiveness of the system was erased as it detected a second appearance of the northern lights. The alarm-determining software has declared no alarms (false or otherwise) for the past one and a half months; last night it issued three. The magnetic field at Boulder, CO was briefly at major storm levels from 5 to 8 p.m. PDT (K index of 7). There were many aurora sighting reports in the northern US; the closest to the aurora monitor's location was a passenger airline en route from Salt Lake City, UT to Pasco, WA.

5 November 1996: Aurora monitor operating again

The aurora monitor is back on the air (145.6 MHz if you live in the Walla Walla valley) and data is flowing again! A big thanks goes to fearless EE professors Rob Frohne and Ralph Stirling who trekked up the mountain to replace crystals, reprogram the monitor with my new code, and tweak the TNC parameters.

24 September 1996:

Finding a suitable location for the completed aurora monitor has not been easy. From the WWC campus, the desired pointing direction (NNE) happens to coincide with the worst light polluter in the valley: the local state penitentiary! Data taken from the roof of Kretschmar Hall (the engineering building) and a location in the Umatilla National Forest indicate that the sky is roughly 30 times brighter on campus, in the the desired direction, and in the filter bandwidth. After searching for a better location with the necessary infrastructure, we placed the aurora monitor on Pike's Peak (a local hilltop) at the broadcast facilities of Positive Life Radio (KGTS). A hefty insulating box was constructed to house the monitor on the side of the tower and hopefully protect the unit from the elements. Data transmission back to campus is accomplished by an amateur radio "packet" link (1200 baud AFSK on 2 meter NBFM). In Kretschmar Hall, the data is fed into a Sun workstation which runs a network daemon that routes everything received on the serial port to a TCP socket. This enables me to tap into the real-time data from anywhere in the world and experiment with alarm algorithms from my new location at the University of Illinois in Urbana-Champaign, IL.

After two weeks of successful operation in early August, we were forced to shut down the link due to conflicts with the local ham population. We were sending packets every 20 seconds on a commonly used packet frequency, which many users found objectionable. Crystals have been ordered for a new, coordinated frequency. They should arrive by mid-October at which time a trip will be made up the mountain to reprogram and restart the unit.

Home - News - FAQ - Tech - Sign up - Data - Gallery

Last modified 15 July 2007 by markrabbitketeuaardvarkorg replace animals with punctuation