Expedition VEGA logo
The Vega 2013 Expedition Greenland started from Gothenburg June 17, 2013, with the 17.5 m long ship Explorer of Sweden. The target is Melville Bay located approximately 105 nautical miles north-northwest of the small village Upernavik on the western coast of Greenland. The expedition is led by Ola Skinnarmo and the scientific components by Martin Jakobsson and Larry Mayer from Center for Coastal and Ocean Mapping, University of New Hampshire, USA.

The Expedition log was written by Martin Jakobsson

The expedition had three main objectives:

1. Locate the wreck of the ship SS Vega that got stuck in the ice and sank in Melville Bay in 1903. SS Vega was used by Adolf Erik Nordenskiöld to sail the Northeast Passage 1878–79. This was the first time the Arctic connection between the Atlantic and Pacific oceans was sailed. 2. Map the seafloor for glaciogenic landforms related to the Greenland Ice Sheet, which has retreated from the outer continental shelf to its present position since the Last Glacial Maximum (LGM) at about 20 ka BP. 3. Map the seafloor to locate possible deep connections and shallow barriers between the outer continental shelf and the inner fjord system. The purpose is to investigate whether or not warmer water could make its way into the fjord system and thereby affect the outlet glaciers of this region.

Map showing the estimated position where SS Vega sunk.

Map showing the estimated position where SS Vega sunk. The background bathymetry is from the International Bathymetric Chart of the Arctic Ocean (IBCAO), which in this region is based mainly on single beam echo soundings acquired when oil companies carried out seismic profiling for exploration purposes. The coastal area is nearly un-charted. The position of SS Vega must be regarded as rather uncertain since it was acquired with sextant. Furthermore, the timing is unknown; is the position representing the location where SS Vega sunk or where she got stuck in the sea ice? Even a small difference in time may imply quite big differences of the wreck’s location due to sea ice drift.

Multibeam sonar: Kongsberg EM2040 200/300/400 kHz, 1x1 deg single head. AUV (Autonomous Underwater Vehicle): Gavia, pressure rated to 500 m and equipped with 600/1200 kHz side-scan from Marine Sonic and GeoSwath Interferometric multibeam/side-scan sonar.

The Gavia AUV is owned by the University of Iceland and the EM2040 multibeam by Stockholm University.

August 15: In Conclusion

Two of the three objectives of the expedition were met beyond expectations; those two were the scientific objectives – mapping of the glaciomorphological aspects of the seafloor and identification of deep passages for shelf waters to enter fjords and impact outlet glaciers. In depth analyses of the acquired data remain. We have also showed that the latest technology in seafloor mapping can successfully be used from a sailing boat, albeit a well outfitted sailing boat for polar expeditions.

Unfortunately the third objective – location of the wreck of Vega was not met. There are several reasons for this:

1. There are many uncertainties associated with the position reported for the location of the wreck of the Vega – uncertainties associated with positioning capabilities of the time (10 km) – uncertainties associated with the nature of the position with respect to the sinking of the vessel (i.e. was this the position where the vessel became stuck in the ice or was this the position where the vessel was holed and sank? To address these uncertainties we attempted to cover an area as large as possible with the multibeam sonar system (320 sq km) – however the depth of the survey area and the remarkably complex topography made direct identification of a wreck from the bathymetric data virtually impossible. Given this topography, which not was known before the expedition, it would have been virtually impossible to use a surface towed side-scan sonar in this region.

2. Follow up with near-bottom AUV-deployed surveys provided backscatter resolution that should be capable of identifying the wreck unambiguously should the sonars encounter it. The AUV is limited however, in its swath coverage and bottom time. This implied that about 10% of the multibeam surveyed area could be covered.

3. The Gavia AUV and Konsgberg EM2040 multibeam sonar results indicate iceberg scours in seafloor as deep as 400 m, thus there is a much greater chance that the wreck has been destroyed by interaction with icebergs. This implies that the chances of finding an intact Vega in this area are rather slim.

The nose cone of Gavia, AUV, with clear signs of hard bottom collisions.
The nose cone of Gavia with clear signs of hard bottom collisions. Photo: Larry Mayer


pil up icon

August 13-14: End of Survey

The weather turned quite bad during the night of the 13th and waves begun to build. This made it very difficult to launch Gavia. Before we had the extremely calm conditions, and we even got relaxed enough during the Gavia missions that some time could be spent on some photo sessions. I posed on an inflatable SUP that was brought along. When I sat down, storm birds came up just next to me to have a look, it was quite a feeling!

Gavia begun to abort mission after mission. She had hit the bottom quite hard in some cases, but there were also some aborts that were difficult to find out the reason behind. Unfortunately Tómas and Arnar eventually found that the aborts were due to that she travelled very slowly because the main motor had taken in salt water. Likely in the very deep areas with depths close to here maximum pressures rating of 500 m. That was it, we had to end the survey and give up the search. However, this was just one day before planned and we had got a lot more data than we ever dreamed of, but no signs of Vega! She will have to remain on the bottom for a while longer without us knowing exactly where.

The weather got quite rough and the trip back to Upernavik was not particularly pleasant. When we arrived midday the 14th, the waves were big even in the harbor.

Martin Jakobsson on a SUP outside Greenland.
On a SUP with curious birds around. Photo: Larry Mayer

pil up icon

August 12: Iceberg giving us trouble
people on a rubber boat infront of an iceberg, Greenland.
Gavia launched in front of a big iceberg in the survey area.

The missions with Gavia continued. They went surprisingly smooth considering the rough bottom topography. The Gavia AUV proved to be impressively reliable in these difficult conditions, but without the careful planning by Tómas and Arnar using the multibeam bathymetry, it would not have worked. The longest missions lasted for close to three hours when they used the better of their two batteries. Each mission provided a section of the bottom mapped with the side scan sonar in high resolution. Larry and I went through all the data in detail in between the missions. The bottom is dominated by crystalline bedrock with abundant boulders and rock formations that occasionally form very ship-like forms.

We had begun with mapping around the position where Vega was assumed to have sunk. There is practically no risk that we would miss her with this high resolution side-scan. But we found no signs of Vega at all at, or around, the assumed sinking position of 74° 23’N 58° 45’W. We proceeded with mapping the targets we identified in the multibeam backscatter and bathymetry. A big iceberg drifted into one of the areas we had planned to map. This caused problems since we saw that it was so deep that Gavia could run into it. The mission was aborted and we waited for the iceberg to move away with the currents from the survey area. It took a while, and the iceberg eventually got stuck on a 200 m deep shoal in the bottom topography. The fact is that we find fresh iceberg scours in there are practically as deep as we can map (>400 m). This certainly is a great risk for Vega, chances are quite large that she has been crushed by icebergs if she not sunk into one of the deep crevasses that could provide some protection.

Seafloor image from a side scan sonar.
The seafloor dominated by bedrock formations as revealed with the side scan sonar imagery acquired with the Marine Sonics side 600 kHz scan sonar mounted on the Gavia.


pil up icon

August 9–11: Gavia survey


We arrived in the survey area midday August 9. The weather had improved a lot; sun and practically completely still compared to the cloudy sky and quite large waves we previously had. Everything looks much nicer than before and life is generally easier on a boat when the sea is calm and sun is shining. The calm weather is also a blessing considering that we have to launch the Gavia AUV from the sailing boat by first bringing her over the reeling and into a rubber boat. Larry and I have worked continuously with data processing. We now have managed to clean all the multibeam bathymetric data. The seafloor mapping has generated a superb view of the seafloor topography. I must admit that I’m quite amazed how well our installation really worked on Explorer of Sweden. No artifacts are visible in the data due to poor alignments of sonars transducers, motion sensor, GPS antennas etc. This boat is excellent for mapping missions. The bottom topography is on the other hand unfortunately much rougher than ever expected! If this bottom topography was known beforehand, I think the hope of finding Vega would have been slim! But no one has mapped the area in detail before. The nearby large glacial troughs to the north did suggest a smoother bottom and that the transition to the rough crystalline bedrock takes place closer to the shore. One thing is for sure, our survey concept was completely right. We simply needed the full bottom information from the multibeam to plan the Gavia missions. It is a risky business to send down the vehicle in this terrain and every steep obstacle must be avoided. To tow a side-scan from a ship over this bottom would simply never have worked! The beauty of the AUV is that it is set to track on a constant height of 25 m over the bottom to be safe, but still be able to acquire high resolution imagery.

Bottom topography.
Bottom topography of the northern part of the survey area. Steep faults run through the area with more than 200 m high cliffs measuring 85° inclinations. Vega is placed at her right scale (white ship) just to get an idea of the object we are looking for in this bottom topography. Find her, but recall that we do not see colors with acoustic imagery!

pil up icon

August 8: Iceland crew arrives with “Gavia”

We returned to Upernavik around lunch time after a smooth transit from the survey area. Tomás and Arnar from Iceland arrived with an Autonomous Underwater Vehicle (AUV) for the second part of the expedition. Their flight was a delayed some hours, but eventually they arrived safely with their “tool”. The AUV is rented to us by Iceland University. It is a “Gavia”, pressure rated to 500 m, equipped with a high resolution (600/1200 kHz) side-scan sonar by Marine Sonic, and a GeoSwath interferomentric bathymetric/side-scan sonar (500 kHz) by Geo Acoustics. Tomás and Arnar began to setup the vehicle at the harbor and eventually made some first test. The first important part is to balance it right for the water density. The setup was swift and we sailed towards the survey area in the evening.

Launching of the Gavia, AUV, from a ship.
Launching of the Gavia in Upernavik harbor.
testing Gavia, AUV, from a rubber boat.
Ballast test of the Gavia.

pil up icon

August 7: “Djävulens Tummefinger”

We decided to go into the nearby village Kuvdlorssuaq, which is located close to the 540 m high top “Djävelens Tommelfinger” (The Devils thumb). This is an amazing rock with a vertical wall that for climbers is a challenging target. The trip to shore was one of the most informative I ever have made during a marine expedition. From the multibeam bathymetry, it became immediately clear that the landscape likely is identical to the area we surveyed offshore! I could recognize the bedding planes in the crystalline and often highly metamorphosed rock.

Explorer of Sweden in Kuvdlorssuaq.
Explorer of Sweden in Kuvdlorssuaq. Photo: Martin Jakobsson
Devils Thumb, 540 meters high rock in Greenland.
The 540 m high Devils Thumb. The overall terrain and bedrock is similar to the surveyed area offshore. This highlights the difficulty of our mission! Photo: Martin Jakobsson
Icebergs outside of Kuvdlorssuaq.
Icebergs outside of Kuvdlorssuaq. Photo: Martin Jakobsson

pil up icon

August 5–6: More mapping

The multibeam system has worked amazingly well. Up until this moment of writing (August 6, 09:00 local time) we have not had more than 15 min downtime. This stop was due to that the main GPS lost positions for a short while and froze-up. The seafloor topography is very rough with numerous bedrock outcrops, hills and crevasses. There are not many flat parts in the area surrounding the estimated position where the wreck of SS Vega may be located. Searching for her in this terrain feels like trying to find a bicycle lost somewhere in a mountain region by flying over the area with a Helicopter during nighttime. No doubt that we do collect fantastic research data! The Greenland Ice Sheet has definitely gone over this area and shaped it. The extraordinary good data quality and high resolution of our mapping data will allow several new analyses to be made after the cruise. So even if we end up not finding SS Vega, two of the three expedition goals will be achieved. Our strategy has been to first do the large overview mapping with the multibeam mounted on Explorer of Sweden and in this data identify potential targets. These targets will subsequently be mapped in detail during the second part of the expedition using an AUV (Autonomous Underwater Vehicle) that is equipped with high frequency side scan sonar. The concept feels right!

pil up icon

August 3–4: Survey continues

Knock on wood, the multibeam works very well. Overall, the sailing boat Explorer of Sweden works great as a mapping vessel! It does not make much noise and there is sufficient space to put up the equipment. The sea is relatively calm which helps the data quality considerably. Now it is just to map map map!

pil up icon

August 2: First survey
Two male scientists in front of the computers.
Larry Mayer and myself logging and post-processing the multibeam data.

The night had been quite rough with a lot of rolling making it a bit hard to sleep. The designed mount of the multibeam is possible to hoist up during transit. It worked really well, in fact the entire installation of our EM2040, 200–400 kHz, Kongsberg multibeam worked extremely well so far. We arrived in the survey area around lunch time. We began with lowering the multibeam into the water. That was quite a bit more difficult than in the harbor due to the remaining swell. But since the wind had calmed down to only about 5 m/s, the swell would soon die out. A couple of survey track were first run for calibration purposes. After this was done, we began our survey. First, abundant disturbances were apparent in the data. We thought initially that this was due to the heave and pitch causing the bow mounted multibeam to even come up above the water level. However, it turned out that this was not the case; it was the ship’s own echo sounding operating at the same frequency of 200 kHz that caused the problem. This echo sounder was switched off, and then really nice data began to come in.

multibeam lines from the area where Vega is estimated to be located.
The first set of multibeam lines from the area where Vega is estimated to be located. The geology of the seafloor in the survey area is quite dynamic with the bedrock out cropping. While the seafloor is spectacular from a geologic point of view, the very dynamic terrain makes it very difficult to find a wreck, in particular of the rather small size of Vega (46 m long).

pil up icon

August 1: Towards the VEGA estimated position

Strong gale force winds in the morning and rain. Not at all good for our so-called patch test, for which calm sea is preferable. We decided not to go out and instead focus on testing other parts of the system, for example the new sound velocity probe from Valeport. I received this probe the days before departure. Therefore, I only had time to make one short test, which I made from my Standup Paddle Board (SUP) in Lake Drevviken. It was indeed a bit scary to paddle out with a very expensive device between my feet on the rather small SUP! Since we will need to lower this probe rather deep (close to it pressure rated maximum depth of 500 m) the fishing winch belonging to our small survey boat RV Skidbladner was brought onboard Exlorer of Sweden. This winch was installed on the aft deck. The test of the SVP probe went fine and so did also the import of the acquired sound velocity profile to the multibeam survey software SIS by Kongsberg. Larry Mayer who will be my multibeam survey colleague arrived 16:45 to Upernavik’s small airport. The wind had calmed down by this time, but there was still a rather large swell remaining. We decided to start our transit towards the survey area and make the patch test there. We left Upernavik at about 19.00.

icebergs off the coast of Upernavik.
Amazing icebergs off the coast of Upernavik. Photo: Martin Jakobsson

pil up icon

July 31: Equipment setup

The weather was quite bad, about 5°and windy. That did not matter so much as this was the “equipment setup day”. We began with assembling the bow mount for the multibeam. Apart from one minor modification due to change of some couplings, it went amazingly smooth. Even smoother than during the trial we made in Gothenburg before departure. As always, it takes quite long time to run all cabling and connect all various sensors (sound velocity sensor, motion sensor, GPS antennas) and make sure they all run as they should and send proper data to the multibeam echo sounder. By late afternoon we had it all running and tried the multibeam in the harbor. Tomorrow is time for a real so called “patch test”. That is when the multibeam is run along some specific survey tracks in order to collect data for calibration of possible offsets between the sensors.

Glenn Mattsing holding on to the multibeam transducer casing.
Glenn Mattsing holding on to the multibeam transducer casing. This casing contains the multibeam transducers (Kongsbeg EM2040, 1°x1°, 200–400 kHz). Photo: Martin Jakobsson
lowered the multibeam to mapping position from the vessel RV Skidbladner
The multibeam lowered to mapping position. During long transits and in areas with sea ice, the sensitive transducers and motion sensor can be hoisted up. The design of the mounting system is taken from our small survey vessel RV Skidbladner. Photo: Martin Jakobsson

pil up icon

July 30: Arrival in Upernavik, Greenland

I Arrived in Upernavik at about 15.00 local time. This is after changing flights in Kangerlussuaq and Ilulissat (Jakobshavn). The latter small village is located next to the extraordinary active ice stream, Jakobshavn breen, which that drains a rather large sector of the western Greenland ice sheet. This ice stream is famous for its fast flow, and for that it has retreated quite significantly over the last decade or so. I was picked up by Ola Skinnarmo and crew at the airport in Ilulissat. His 17.5 m long sailboat Explorer of Sweden was docked in the small harbor outside a luxury cruise ship named Lady MII. This approximately 50 m long cruise ship is obviously solidly constructed as the owner and skipper have decided to embark on a journey through the North West Passage this summer.
Upernavik is a very small fishing village. The small wooden houses are neatly lined up along the steep shore line and painted in different eye catching colors. Later tonight we will begin unpacking the multibeam equipment, which has gone with the sailing boat from Gothenburg.

Signs outside the Airport in Kangerlussuaq.
Airport in Kangerlussuaq. Photo: Martin Jakobsson
Icebergs flowing out towards the open ocean in Illulisat.
Seagull lifting from an iceberg in Upernavik.
Seagull lifting from an iceberg in Upernavik. Photo: Martin Jakobsson
The cruise ship Lady MII in Upernavik.
The cruise ship Lady MII in Upernavik.

pil up icon