Data Repeatability and 3D View of the Snowpack
Before we ship each Snow Scope Probe to customers, we do a final device test on snow to ensure quality. Recently, before shipping out a batch of probes, I took the opportunity to compare data from this test across 5 probes and see if I could generate a 3D view of the snowpack.
This blog post gets a bit into the weeds, so if you are unfamiliar with the Snow Scope Probe, you may want to read some of the content on our Scope page and Learn page.
To start, I collected 25 total Scope profiles - 5 with each probe in a grid pattern (spaced about 10-15cm apart) as shown below.
One of these 25 profiles is shown below, compared to a manual hand hardness profile collected in the same spot. The Snow Scope Probe matched the hand hardness profile very well, and picked up the major stratigraphy of the snowpack.
This Scope Profile is shown on our automatically generated “Hand Hardness Equivalent Scale” - for more details on what this means and how it correlates to manual profiles - check out our learn page.
I then plotted each of the 25 profiles in a grid corresponding to the probe location. Small thumbnails of each profile are shown below superimposed on their approximate locations. You can see that the profiles were very repeatable across the grid - it’s difficult to tell from this view that the profiles are even different!
The 3 blank profiles were errors generated from the Snow Scope - occasionally the Snow Scope will recognize that the collected data was not high enough quality to generate an accurate profile and will indicate an error.
While this previous visual is great to see all the profiles in the standard hardness vs depth view that we are used to looking at from hardness profiles, it is a bit difficult to compare profiles side by side.
Below, I generated a heatmap showing a single horizontal transect of 5 profiles. The colors in the heatmap correspond to different snow hardnesses. In this plot, you can better visualize these hardnesses as snow layers - imagine viewing this as the side of a pit wall, and the different colors correspond to different layers across the pit face.
While it looks like the layers change a bit across the face of the profile, it is important to note that the total depth of this plot is only 50cm, so small changes in height (+/- a few cm) are visible here — these 5 profiles are all very similar.
These horizontal transects of 5 profiles each can be generated for each of the 5 rows of profiles. You can see similar patterns in each of the 5 transects, with the obvious harder layer about halfway down slightly deeper towards the center of each transect.
Again, the 3 black bars correspond to error profiles (missing data).
Finally, Those 5 horizontal transects can be integrated with vertical transects to make a 3D representation of the snowpack under the grid. This image shows one example of what that could look like. The “cutout” in the 3D snowpack corresponds to the area below the orange line in the image. The color marked hardnesses align very well, and are easily identifiable as “layers”.
While this example isn’t immediately very practical, it is a good demonstrator of the repeatability of the Snow Scope probe, and of examples of visualizations you could create with the data.
We’re excited to see what you can do with the Snow Scope data you collect - if you come up with anything cool or have better ideas on how to visualize the snowpack in 3D - let us know!
