Friday, August 8, 2014


The inland ice of Greenland is definitely a cold place. And one of the first questions many people ask when talking about our expeditions is on how cold it actually was. Yet the answer is far from obvious. And even though we almost always carry a thermometer, we rarely ever have a look at it. And when we take a measurement, its often because the weather either feels remarkably cool or warm. But what about the average? The lowest temperature at night? The maximum temperature during daytime?

Wouldn't it be nice to have a continuous recording?

Cold chain surveillance

At first sight, there is a huge choice of temperature loggers available on the market. But a closer look immediately rules out most offers by the demands imposed by polar kiting expeditions:

- temperatures down to -40 degrees Celsius
- autonomy (memory and battery) of more than 2 months
- shock-proof
- small, lightweight package
- easy connectivity, preferably via USB port
- a data format allowing simple extraction of the data for further treatment and presentation
- resolution better than 1 degree Celsius
- affordable price

To my surprise, there is one industrial application meeting exactly all the above specifications: temperature loggers for cold chain surveillance of frozen food and medicals. After some research on the web, I settled on the 3M TL30 (manufacturers website) and bought two units for a test during Wings Over Greenland II.

We had two 3M TL30 temperature loggers. Originally they are intended for
cold chain surveillance. They are small, lightweight (30 g)have an autonomy
of up to one year and are  not afraid of cold temperatures. We protected them
in plastic bags against  moisture and white (spectacle) pouches from
absorption of sunlight.
In order to avoid malfunction due to condensation or accidental immersion (sea ice, surface melt puddles) the units were sealed in tiny plastic zip-lock bags. Professional outside thermometers have sophisticated ventilated white housings to avoid absorption of direct sunlight. As a first approximation, I packed our loggers in small, white spectacle pouches, that I found at the french  sport store "Decathlon".

How and where to measure?

As with every new device, it takes some time to figure out how to use it, and in this case also to define some routine. And so, the units sadly spent the first part of the trip waiting  inside an electronics bag, that rarely ever left the sled. In the meanwhile we observed, that the pulks and our new yellow/red tent heated up considerably during daytime and got interested in temperatures again. Only then did I decide the units should be outside the pulks during progression, and one outside and one inside the tent at the camps respectively. From camp 16 on, unit 1 was placed outside the tent, and from camp 20 on, unit 2 inside the inner tent pocket close to the entrance. During progression both were attached to a small backpack that I always wore to carry a few fragile things. 

At the camps, logger number 1 left outside, close to the tent entrance. It
was covered with a few centimetres of snow to avoid direct illumination.

At the camps, logger number 2 was stowed away inside one of the tent
pockets in order to get an estimation of the inside temperature. It would
have been better to suspend it somewhere in the middle, to  avoid direct
heating through the tent wall when the sun was on this side of the tent.

During progression both loggers were attached to my backpack. Next time
I would avoid attaching them to a black surface.

These were practical choices and all three locations have their shortcomings:

The backpack is black and tends to heat up, at least when its not exposed to the wind. This becomes particularly apparent in little marked spikes in the outside temperature, just before sensor 1 was put onto the snow during camp. But also during the days there are spikes that are too sharp to be realistic. Its likely that at least the maximum temperatures are exaggerated.

Because of the wind, logger 1 was on the snow, attached to one of the tent pegs and covered with a few centimetres of snow at the camps. The snow surface temperature does not necessarily follow the air temperature.

The inside thermometer was always in the inner pocket close to the entrance. This may significantly exaggerate the reading depending on which side of the tent is exposed to the sun, and it would certainly have been more wise to suspend it somewhere in the middle of the tent. 

But once I realised all of these issues, we had already completed a substantial part of the trip, and I thought it better not to change protocol once more mid-way. 

I thought its worth sharing these experiences anyway.

How cold?

The following image shows the temperature trace of  the "outside" logger 1 for the entire expedition. The black vertical line delimits the first part, where the logger was still inside the pulk, from the time, when it was systematically placed outside next the tent entrance or attached to my backpack. The large relative shift in the data makes the problem directly apparent. 

The first striking observation are the large oscillations due to the diurnal variations of the temperature. The temperatures are strongly correlated with the height of the sun above the horizon. A closer look shows, that the average temperature is globally decreasing during the first half of the trip before rising again in the second half. This correlates well with the latitude, as the start and end of the trip was the southernmost point, and we reached the turning point at latitude 81 N on 19/05. At the same time, the diurnal variations appear to decrease while heading north and to slightly increase again on the way back south. The effect is however barely significant, as the days were still getting longer and cloud cover and the spikes in the data play an important role.

Temperature readings from logger number 1 (outside). Click to enlarge.

To make this a bit more visible, the next graph shows the evolution of the daily average temperatures, along with the minimum and maximum values. (The minima and maxima were determined using a moving average over 1 hour, in order to smooth out spikes.)

Average temperatures (black), along with the minima and maxima. Click to

We have passed two automatic weather stations on the icecap (Nasa-U: news "Sport meets Science" and Humboldt-Gl: news "Humboldt - turn right".). It will be interesting to compare the temperatures for these days, as soon as this years data from the weather stations becomes public (supposedly soon).

Cosy interior

The following graph shows the temperature inside the tent together with the outside temperature for one week starting on saturday 17/05. Its impressive how much the tent (Helsport Svalbard 5 Camp)  with its yellow and red tissue heats up in direct sunlight.

Temperature readings from both loggers for the week from  May 17. 
Click to enlarge.

The image below of our 'analog' freezer thermometer and Mika, taken at the very beginning of the trip shows, that sometimes it can get really hot inside (news "Sauna")!

The end of a polar myth: in direct sunlight, the temperatures inside the tent 
can rise far above zero.

Lessons learned

Temperature loggers for cold chain surveillance are a simple, way to document the conditions during a polar expedition. They are intrinsically robust, have a long battery life and are made for continuous use at low temperatures. It turns out, that getting correct temperature readings is anything but an easy task! It is worth to be carful with the packing and  and to think about the placement of the loggers. I would do it differently next time. Pulks, bags and tents heat up far above outside temperatures. Speaking of cold chain surveillance: it is far from safe to assume that the food is always below freezing temperatures during a polar expedition.

The software required to read the logger is only available for windows, and the battery is not user replaceable.

If you are interested in using this type of logger during an expedition, drop me a line, I would be glad to help and to share experience - properly done, learning from our mistakes, carrying loggers is certainly worth the effort!

Thanks to Florian for helping me out with a windows computer to read the data loggers - I do not own such a thing anymore.

No comments:

Post a Comment