What is a “Smart Probe”
"Smart Probe" has been a term in our industry for many years but the technology and value of these devices is still not well understood. In this blog post, we'll go back to the basics and talk about what a smart probe is and what value they can add to snow measurement.
What is a "Smart Probe"?
A "Smart Probe" is a device that is probed into the snow and measures one or more qualities of the snow. Unlike your avalanche rescue probe, a smart probe's main function is data collection, not rescue. Most smart probes measure the hardness of the snow or more accurately the penetration resistance of the snow as you push the probe through each snow layer. Probes that measure penetration resistance are also called penetrometers. There are non-smart penetrometers out there such as your ski pole when you flip it over and "feel" for layers as you push it into the snow. Another slightly more scientific example is the RAM Penetrometer that relies on weights and recording measurements to discern snow structure. Recently, people have been adding electronics and sensors to probes to try to more accurately and easily measure the same thing. The electronics and sensors in these new digital penetrometers has led our industry to distinguish them from their predecessors by adding the label "Smart".
Some Examples:
How Does it work?
Snow structure is a core part of snowpack assessment for avalanche danger. Snow structure is typically measured using Hand Hardness profiles. Hand hardness profiles fundamentally measure the same thing as penetrometers and smart probes: penetration resistance. Penetration resistance is essentially the pressure at which your measuring device penetrates each layer of the snow. Without getting too technical, pressure is force divided by area. When we do a hand hardness profile, we keep the penetration force constant (or try our best to) and vary the area of the object penetrating the snow(e.g. fists, fingers, etc.) to determine the pressure or penetration resistance of different layers. A penetrometer measures the same thing slightly differently: the area or the probe stays constant while the force to penetrate each layer varies. Analog penetrometers measure this force with calibrated weights in the case of the RAM and "feel" in the ski pole test. A Smart probe measures this varying force with a sensor.
So while penetrometers, smart probes, and hand hardness are all trying to measure the same thing, why do the snow profiles from each all look so different? The mechanics of how we measure ends up affecting how the data look. Hand hardness profiles typically have very well-defined layers, often with a single hardness value for the entire layer. The reason for this is that we are limited by our hand area to measure layer size, so the smallest fist hardness layer you can measure is the height of your fist. Because of this size constraint, you often only get one or two measurements per layer, which further contributes to the clean, simple look of hand hardness profiles. Smart probes, on the other hand, typically have a very small measurement surface and take dozens or hundreds of measurements per "layer". This design often results in many measurements within a single "layer" and explains why smart probe profiles, including those generated by the Snow Scope, looks much more detailed without relatively large, well-defined layers. I have put layer in quotations to draw attention to the fact that most smart probes do not understand what constitutes a layer boundary, and that is still up to a human to interpret from the data.
Smart Probe Data vs Hand Hardness Data: Same Structure, different representation
Why are "Smart Probes" important?
The big idea behind a smart probe is that it can do the job of measuring snow structure faster and more accurately than a human can. Now before we sound the robot apocalypse alarms, hear me out. Computers do some things much better than humans, recording measurements and doing simple math quickly is definitely one of those things. Interpreting those measurements is a whole different question and may deserve its own blog post, or probably its own research study. To put things in perspective, the Snow Scope Probe can give you a snow profile in under 10 seconds from taking out the device, much faster than your typical hand hardness profile. This allows tens if not hundreds of data points to be collected in the time it would take to complete a single hand hardness profile, allowing spatial variability differences to be captured that would be missed with a single snowpit.
On the accuracy side, hand hardness pits are full of opportunities for human and procedural error. Going back to our equation for penetration resistance/pressure, the error or bias in hand hardness profiles comes from error in force and area. If you do not use the exact same force in your hand hardness "pushes", you are introducing error. Furthermore, everyone's hands/gloves are not the same size, and this introduces bias in the area part of that equation. Finally, there is error in simplification of the snowpack into layers; each individual recording a snow profile may define the number and location of layer boundaries differently. These sources of error make hand hardness profiles useful for general snow structure measurements, but comparison or data analysis are often impossible due to human error and bias in hand hardness pits. The figure below shows how "one hardness to one person might feel like another hardness to someone else.
Pressure Ranges for different hand hardness measurements from different people. See the overlapping ranges? That means that hand hardness isn’t giving most practitioners a repeatable measurement of the snowpack! Adapted from “Quantification of the hand hardness test by P.Holler, R. Fromm, Annals of Glaciology, Volume 51, 2010”
Smart probes solve this problem by taking human and procedural error out of the equation. Smart probes are not infallible and can be prone to measurement error. Measurement error is anything that causes a sensor to report a value different from the value being measured. For example, a sensor designed to measure soft snow may be maxed out if used in harder snow, or if a particular sensor that measures snow depth does not work in direct sunlight, these can introduce measurement error into the profile produced by a smart probe. I would argue that this measurement error is much easier to deal with than human bias because it is often well documented and bounded. The objectivity of sensors and computers over humans in measuring the snow is where the real value of smart probes lies.
Separate Scope measurements show the repeatability or objectivity of smart probe measurements
This value is largely still unrealized due to the fact there has not been a smart probe accurate and convenient enough to be widely adopted. Most smart probes either struggle with technical challenges around accuracy and reliability or, if they have solved those problems, it likely requires a PhD and funding to obtain. The benefits to the snow science and safety industries of a well adopted snow hardness measurement device are numerous. More objective snow data will engender better research and further our understanding of snow science, which will result in better models and avalanche forecasting for use by industry pros.
Here at Propagation Labs, we are thinking about these problems every day, and are acutely aware of the challenges in creating a device worthy of adoption by our industry. An industry which has seen many failed attempts and may be harboring some healthy skepticism toward smart probes. While biased, we feel that our smart probe, the Snow Scope, has the potential to convince the skeptics and begin to unlock some of the untapped value “Smart Probes” have to offer.
