By Chris Polashenski and Simon Filhol

It has been a very busy time for us the last few days. After about 8 days of prep work, our aircraft arrived in beautiful weather. Our pilot, Paul, and Dr. Chris Larsen started flying airborne LiDAR immediately, and as will be seen, collecting enormous amounts of data. Meanwhile the rest of us were working feverishly on the ground to collect data against which the airborne LiDAR could be compared.

A schematic of the whole campaign appears in Figure 1. Note the snow scientist labeled ‘E”. He is using a device invented by CRREL called a MagnaProbe, which measures snow depth and a GPS position at the same time (Fig. 2). At peak, we have had 8 of these devices out at once, measuring snow depth in a variety of areas. Nonetheless, despite this herculean effort, the “probers” efforts pale in comparison to those of the LiDARians, as explained by the two young scientists doing the ground-based LiDAR who take up the story now.

Figure 1 – A schematic of the SnowStar 2012 Campaign by Matthew Sturm, compare to the photo in Figure 18 to see how we did!

Figure 1 – A schematic of the SnowStar 2012 Campaign by Matthew Sturm, compare to the photo in Figure 18 to see how we did!

Figure 2 – Sveta collects snow depth measurements with a Magnaprobe

Figure 2 – Sveta collects snow depth measurements with a Magnaprobe

“N” is a very important concept to scientists here on the SnowStar 2012 campaign. To us, a person’s “N” is a measure of his/her gallantry, the quality of their character, and their general worthiness as a scientist. “N” is the number of data points a person collects in the day and a high “N” is a sure sign of a lion-hearted individual.

The teams measuring snow depths with the automatic snow depth probes (see Fig. 1 and 2) are engaged in a stiff competition to show their mettle with ever higher “N”s each day. The highest reported N in the campaign so far has been about 4000 in a day. One legendary effort, still discussed with awe by snow scientists, resulted in an “N” that approached 11,000 over the course of a very, very long day (and on sea ice so the snow was thin). So far in this campaign no one has come near that figure due to the large field area, the deeper snow, and the long commute to each site.

The ground-based scanning LiDAR collects that awe-inspiring 11,000 points in less than half a second. Hence we (the ground-based LiDAR team) have been averaging between 20 and 30 million points at each site, and we’ve knocked out as many as four sites in a day. In the world of snow measurements, LiDAR is to a snow probe as dynamite and steam drills were to hammers and chisels in the old railroad days. And while we shed a tear for all the John Henrys out there still breaking stones by hand, we have not been shy when bragging about how our LiDAR could handily best the snow prober’s ‘N’ any* day of the week. (*Windy, foggy, or snowy days excepted – see Figure 3).

In all seriousness, Big “N” means that LiDAR holds real promise for revolutionizing our work by making it possible to measure snow with less manpower, more accurately, and over much larger areas, but we still have to work out the kinks in the system. This program is all about just that – figuring out how we can better measure snow and comparing new methods like LiDAR with the old fashioned ways.

Figure 3 – LiDAR fails as the snow flies.

Figure 3 – LiDAR fails as the snow flies.

LiDAR (which stands for Light Distance And Ranging) is a technique that uses a laser rangefinder to measure surfaces. Basically, you fire a pulse of laser light at a surface, and with an extraordinarily precise stopwatch, time how long it takes to get back to you. Using the speed of light, we can then use the time-of-flight to calculate the distance that the pulse traveled, divide by two, and have the range to the surface.

If we know exactly what direction the pulse was ‘fired’ at, we can use this range to calculate the location of the snow surface as a specific point in space. By firing the laser over and over again in slightly different directions, it is possible to collect many samples of the snow surface. Putting all of these point measurements together gives the LiDAR’s product: a 3-D map of the snow surface topography.

The 3D surface maps that our LiDAR produces are incredibly high resolution (at least a few hundred points per m2) and super accurate (within about 1 cm) – over areas the size of a couple football fields. Aside from being scientifically useful, these 3D maps are just plain awesome to play with. Check out this video flying around in the 3D world created from scans we took at a site in Fairbanks just before this trip. Also in the screenshot shown in Figure 4 from this campaign, note that small twigs and even the ptarmigan tracks in the snow that have been picked up by the LiDAR.

Figure 4 – A screenshot of LiDAR data collected on this trip, note the snowmobile track and footprints across the bottom of the image, and the ptarmigan tracks wandering around the snow surface.

Figure 4 – A screenshot of LiDAR data collected on this trip, note the snowmobile track and footprints across the bottom of the image, and the ptarmigan tracks wandering around the snow surface.

The LiDAR that we are using is considered a terrestrial or ground-based LiDAR, complicated speak for the type of LiDAR that you use from a tripod sitting on the ground (Figure 5). Because the LiDAR works by hitting the surface with its laser pulses, it can only scan what it can see. Since the LiDAR is positioned about the height of a person’s eyes off the ground, the back side of even small hills within the scan area may be out of sight from the LiDAR’s perspective and therefore missed in the scans.

To fill these in, we set up the LiDAR in different positions around an area of interest, making scans first of one side of hills, then of the other, and overlay the scans to build a full surface. In order to overlay the scans accurately, we put targets out around the site and use the scanner to locate them from each position (Figure 6). The data collected at each site is then rotated and shifted until all the targets line up, allowing us to very precisely tie the scans collected from different viewpoints together into a single surface map.

Figure 5 – Simon operating the ground based LiDAR

Figure 5 – Simon operating the ground based LiDAR

Figure 6 – Scanning a reflector target with the LiDAR to overlay the scans. The green dot is the LiDAR laser.

Figure 6 – Scanning a reflector target with the LiDAR to overlay the scans. The green dot is the LiDAR laser.

One of the key measures of snow that we are after is simply its depth. The LiDAR does not, however, measure snow depth, just the surface position. To calculate the snow depth, we will have to come back in June after the snowmelt and re-LiDAR the same locations to create a second surface. Subtracting the ground surface from the snow surface will give us a map of the snow depth with a really big “N” and accuracy that is likely as good or better than the snow probes, but only if we can come back to exactly the same spot. Accurate GPS positioning is crucial to this process.

Despite having this new tool, Simon and I have been working some rather long days (Figure 7). This is proof in my mind that scientists aren’t really all that smart after all. Prior to now we had a tool that would take 11,000 points in a day with a lot of hard work. Now we have a device that takes 30,000 points a second, and instead of letting it run for a few seconds and taking the rest of the week off, we still scan all day and well into the night, all in the pursuit of bigger ‘N’.

LiDAR scanning isn’t such bad work though. During the data collection process, there is a period of about 20 minutes at each scan position when the scanner is acquiring data at the speed of light, but there really isn’t a whole lot for us to do except try to stay out of the way and not block the scanner’s view. Some might view this free time on the tundra as little more than a good way to get cold. Instead, it is our favorite part of the process.

Simon notes that “you have to imagine being in middle of this gigantic white landscape with few hills around and an astonishing mountain range in the background, few herds of caribou grazing, ptarmigan flying from bush to bush, a couple of wolves cruising, a lonely musk ox we’ve named Uncle Jacquis Alfred Dalton (for many reasons we will not mention here…) and two young snow scientists, one American and the other French.

In this situation we usually take one of three options: admire the surrounding nature (Figures 9-13) and explore the wonder of our favorite element (snow of course), discuss and argue on all sorts of topics with a preference for cultural contradictions between our respective heritages, and finally let our imaginations create little games or warm up dances (Figure 8). Before you know it, work calls back and we move to a different site position."

Figure 7 – The sun sets as Simon gets the GPS base station set up to start another site, committing us to another 3-4 hours of work.

Figure 7 – The sun sets as Simon gets the GPS base station set up to start another site, committing us to another 3-4 hours of work.

Figure 8 – Simon dances through LiDAR downtime.

Figure 8 – Simon dances through LiDAR downtime.

Figure 9 – Not a bad view from LiDAR site 9.

Figure 9 – Not a bad view from LiDAR site 9.

Figure 10 – A fox looking a bit guilty with a Ptarmigan feather in his mouth visits the LiDARians.

Figure 10 – A fox looking a bit guilty with a Ptarmigan feather in his mouth visits the LiDARians.

Figure 11 – Uncle Jacquis Alfred Dalton demonstrates the effectiveness of long fur for showing wind direction.

Figure 11 – Uncle Jacquis Alfred Dalton demonstrates the effectiveness of long fur for showing wind direction.

Figure 12 – A pair of wolves investigate the LiDAR team just north of the Brooks Range.

Figure 12 – A pair of wolves investigate the LiDAR team just north of the Brooks Range.

Figure 13 – The Northern Lights are a good reward for working late.

Figure 13 – The Northern Lights are a good reward for working late.

For all of our bravado about the power of our ground-based LiDAR scanner for collecting big N, there is another type of LiDAR out there that makes our results pale in comparison. The Airborne LiDAR! It is the carpet-bomb of all snow measurements, if it can be made to work. An airborne LiDAR unit is flown looking down out of the belly of a plane and runs continuously as the plane flies, collecting 10-100 times as many returns in a day as our best ground-based efforts.

There is a trade-off, however: the airborne LiDAR has a lower accuracy; ~10cm rather than ~1cm. Nonetheless, yesterday the ground LiDAR team’s ‘N’ was humbled by this device in just 4 hours, when the airborne team returned over 300 million points before lunch, to our 65 million all day. Adding to our jealousy, the Airborne team also gets to trade our already pretty awesome snowmobiles for an absolutely amazing 1000hp, turbine-powered single otter plane (Figure 14).

Figure 14 – The airborne team takes to the skies

Figure 14 – The airborne team takes to the skies

Figure 15 – Controlling the airborne LiDAR from the cockpit.

Figure 15 – Controlling the airborne LiDAR from the cockpit.

Figure 16 – The airborne team passes by the ground team

Figure 16 – The airborne team passes by the ground team

The struggle for data collection superiority will continue for a few more days, but for now we are all excited to say that good weather and a stellar team is leading to ‘Big N’ for everyone on the trip. Better still, our efforts to coordinate the different teams (Figure 17) have been going quite well. Our goal of collecting a huge dataset of snow measurements over the same areas using a number of different techniques is really coming together (Figure 18). We will have some mighty data processing to do when we get home!

Figure 17 – Coordinating the teams

Figure 17 – Coordinating the teams

Figure 18 – The whole campaign coming together.

Figure 18 – The whole campaign coming together.

About the Authors:

Chris Polashenski has been working to better understand snow and ice in the Arctic since 2005. He graduated with his PhD in June 2011 and now works as a researcher at the U.S. Army Cold Regions Laboratory developing new methods to study snow and sea ice. He is a native of Pennsylvania, but currently lives in Hanover, NH with his partner Norah, chickens, rabbits, and a free spirited beagle named Tracks. He can be reached at christopher.m.polashenski@usace.army.mil.

Simon Filhol has previously worked in the Arctic in Norway, Svalbard, and other parts of Alaska. He hails from Chamoix, France and is currently working on a PhD in geology and geophysics at the University of Alaska Fairbanks with Matthew Sturm. He can be reached at svfilhol@alaska.edu.

 

 

===========

 

 

Previously in this series:

Alaskan North Slope Snow LiDAR Campaign: SnowSTAR-2012

SnowSTAR-2012: Hoars and Drifters

SnowSTAR-2012: Questionable Monuments and Widespread Cratering

SnowSTAR-2012: Big “N” – The pursuit of snow data and high honor