Hessdalen

What the 1984 report is

“Project Hessdalen 1984 - Final Technical Report”, by Erling Strand, is the record of a winter field campaign in the valley. It is signed “Erling Strand (sign.) 5 January 1985” and published on the project’s website with its appendices: a summary of every report, the radar recordings, the spectrum analyser and magnetometer readings, and photographs.

It is not a state document, and it says why: “As no official institutes with governmental support seemed to care about these unknown lights, five persons started their own project.” The project had a working committee of five, an advisory committee, and a local expert group “consisting of people from Norwegian Defence Research Establishment (NDRE), the University of Oslo and the University of Bergen”. NDRE helped with “planning the field investigation, getting instruments, and discussion of the results”. The field work was done by volunteers “free of charge”, and the project was supported by “UFO-Norway”, “UFO-Sweden” and “Föreningen för psykobiofysik”.

The report’s background: “In December 1981, unknown lights suddenly started to show up.” They were seen mostly in the evening and at night, more in winter than in summer; “At the most, it could be seen about four times a day.” In the spring of 1983 they became much rarer, and in the summer of 1983 there were no reports.

The campaign

The project “was born on 3 June 1983”. A seismograph was installed in the valley on 24 October 1983, and in January 1984 an information bulletin with a report form went to 3,300 households. “The work in the field, with all the instruments, started on 21 January 1984, and ended on 26 February.” About 22 people came for the test weekend; between it and the main period, 11 to 26 February, “only one or two people were on duty” at times.

The headquarters was a caravan at Aspåskjølen, “with most of the instruments in it”; in the main period the field stations were at Hersjøen and Litlfjellet. The instruments (section 2): cameras with gratings for spectra, the seismograph, an “Atlas 2000” radar with a wavelength of 3 cm and a maximum range of 33 km, a spectrum analyser sweeping from 150 kHz to 1250 MHz, a fluxgate magnetometer, a 633 nm laser, three Geiger counters and two infrared viewers.

The valley, in the report’s words, lies “south-east of Trondheim, about 30 km northwest of the town Røros”; it is “about 12 km long, and about 150 people live there”. The pin is OpenStreetMap’s point for Hessdalen (node 249179423), rounded by us to 62.79° N, 11.19° E; it marks the place, not a position of any light or station.

What the instruments recorded

Reports. “During the project period in the field, from 21 January to 26 February, we obtained 188 reports on different lights.” Each was scored from G1 to G10 for the quality of the report and from F1 to F10 for how hard it was to find a known explanation: “F1 marks with a hundred percent certainty that the light is known; i.e. an aeroplane.” The scoring was “a subjective decision, but it is made by two persons.” The report’s table puts 135 reports, F1 to F4, under “A known light” and 53, F5 to F10, under “The Hessdalen Phenomenon”, with the threshold the author chose: “A report which has got an F5 or higher might be the Hessdals-phenomenon.” The 188 include “the reports that were written on planes too; 86.” Six reports scored F9 or higher, “too little to draw any conclusion about the nature of the light”.

Radar. “Totally, we have 36 radar recordings. Three of those were probably also seen as lights. All the others were not observed visually.” Nine times a reflection could be followed “travelling on a nearly straight line”; photographs taken where the radar showed something “didn’t show anything unusual”. The radar was mostly set to a range of 5.5 km. Two reflections were photographed on the screen, and an NDRE radar expert said of them: “If this isn’t a reflection of a solid object, but only gas in the air, the gas has to be locally and strongly ionised. Otherwise, it wouldn’t give such a strong reflection”. For one of the three coincidences, on 27 January, the report works out from two echoes 2.4 seconds and about 20 km apart “a mean speed of about 8000 to 9000 m/s”.

Spectra. Of seven films taken through gratings, four frames showed some kind of spectrum. The two strong ones were of a light scored F5: “With other words; it might be a known light.” A weak one, analysed by computer at the University of Oslo, was “too weak to say what kind of spectrum it is”. The report: “these pictures cannot be used as a proof that the Hessdals-phenomenon has a continuous spectrum.”

Other instruments. The seismograph recorded no local seismic activity: “Until now, we have not found any connection between the phenomenon and movements in the ground.” The spectrum analyser recorded “spikes” about 80 MHz apart, never at the same time as a light: “A natural explanation of these recordings, is noise from the radar”, although the radar screen showed no noise at those times. The magnetometer showed “a correlation of about 40% with the pulsation-registrations” over four days, with the caveat that “there is a possibility that there was a coincidence”. A laser pointed at a flashing light nine times was followed by a change in the flashing eight times. The Geiger counters and the infrared viewers recorded nothing, and no light came within 1 km of the counters.

The project’s website now says: “During the 1984 field study, 53 phenomena were documented with instruments ranging from cameras and radar to magnetometers and infrared viewers” (home page, read on 16 October 2026). In the report, 53 is the number of reports its authors scored F5 or higher, and three radar recordings coincided with a light.

What the report says went wrong

Section 4 of the report sets out “what we should have done differently”:

• People. “We did not have a continuous watch on all stations”, and the radar and spectrum analyser screens were not watched all the time, “mainly due to lack of personnel”.
• Planes. Without a direct line to the radar station covering the district, the planes could not be picked out, which left many reports between F2 and F8.
• Spectra. “Many of the spectral films had too much noise; higher order spectra from known lights”, for want of a special screen on the cameras; exposures were sometimes too short.
• Magnetometer. There was no pulsation magnetometer, and the fast recorder was there only from 11 to 14 February.
• Photographs. “Some of the best pictures did not agree completely with the report.”
• Failures. “Many times during the project period, the cameras failed. This happened mostly when we needed them most.” Other instruments failed “mainly because of bad power connection”.

The conclusion stays within those limits: “We have not found out what this phenomenon is. That could hardly be expected either. But we know that the phenomenon, whatever it is, can be measured. Beside the light measurements, it can be ‘measured’ by radar and laser.”

The station since 1998

The project describes itself on its website as “A non-profit volunteer Citizen Science organization”, “Established in 1983 to study the phenomena and in 1998 The Blue Box was installed as the world’s first automated Anomaly Detection System.” It streams the station’s cameras live and puts their pictures in a public folder (home page, read on 16 October 2026).

The same page reports an observation of 1 September 2026, made during a field trip: “For the first time since the 2024 camera upgrade, a witness report could be verified through camera footage. A light has been seen (witness report below) which also appears (very small) on one of our cameras.”

Why this scores where it does

Witnesses: 3. Valley residents and volunteer observers, 188 reports in five weeks. Not higher: the scoring was subjective by the report’s own account, and most reports were of known lights.

Radar: 3. 36 recordings with dates and times, two photographs of the screen assessed at NDRE, and three coincidences with a light. Not higher: 33 recordings had no light, the screen was not watched all the time, and the radar was mostly set to 5.5 km.

Optics: 2. Photographs of lights and four spectral frames. The spectra settle nothing, and the cameras failed “mostly when we needed them most”.

Physical trace: 0. None.

Multi-channel: 3. Eyes, cameras, radar, laser and magnetometer at the same place and time; radar and eye met three times.

Documentation: 4. A dated report whose appendices list every report and recording. Not 5: it is the project’s own report, and some of its best photographs did not agree with the reports.

Anomaly: 3. A few lights the project could not explain and strong radar reflections without a visible light; the report itself says it did not find out what the phenomenon is.

Data today: 3. The report and its appendices can be read, and the project puts the station’s camera pictures in a public folder. The 1984 magnetometer chart, “several meters long”, is not in the report, and the radar is known from a list, drawings and two photographs.

The overall is 2.6: class C, material gaps.

What would settle it

• A radar recording and a camera record of the same light at the same time, with its distance, which the 1984 campaign got three times at most.
• Spectra of a light the project could not explain; the report hoped “that we will succeed in taking a spectral picture of an F10-light next time”.
• What the 2024 paper itself calls for: “Further statistical examination of long-term geophysical records will be necessary for assuming a further accurate interpretation.”
• The station’s records, published with the times of the sightings they are matched against.