GPS Receiver Noise Can Be Used To Detect Snow Depth

cremeglace writes "Scientists at the University of Colorado at Boulder have found a use for GPS besides finding restaurants or the occasional road-that-doesn't-exist: it can be used to measure snow depth. The new technique, which takes advantage of distortions of the GPS signal after it reflects off the snowpack, may potentially improve weather forecasts by allowing meteorologists to track snowfall patterns. ScienceNOW has the story, which one geophysicist describes as 'a classical case of one person's noise becoming another person's signal.'"

Comments

Re:Simpler tool

[Thornburg]

TFA is talking about using an existing network of 1100 GPS receivers currently tracking plate tectonics to also track snowfall, without any additional equipment and without interfering with their current operation.

This not about using your Garmin to find out how much snow is in your front yard.

Re:So, we've discovered

[Wolvenhaven]

You can't stop the signal Mal, you can never stop the signal

Re:So, we've discovered

[PeterBrett]

satellite radar altimetry.

Of course, it's accidental radar altimetry, rather than a dedicated instrument. Neat hack.

Um, no. It's not like SRA, apart from the use of a satellite, RF radiation, and the measurement of a distance.

Main differences:

• This system is bistatic; SRA is monostatic.
• This system uses an (almost) isotropic antenna to collect radiation from pretty much everywhere; SRA uses a high gain antenna to survey only a very narrow target swathe.
• This system measures the effective speed of light in a multipath environment, assuming the multipath reflectors are at fixed distances; SRA measures the distance of multipath reflectors, assuming the effective speed of light is fixed.

Disclaimer: I work on satellite synthetic aperture radar, which is different again, and my knowledge of SRA isn't comprehensive.

Your might think it's unimportant.

[plopez]

All jokes aside, the western US and Canada are completely reliant on snow pack for water supply. No snow, and you have severe drought. Knowing what is happening with snow pack is a huge issue there and in may other places in the world.

http://www.agu.org/pubs/crossref/2007/2006WR005653.shtml

Re:Your might think it's unimportant.

[dschl]

We already know, as there are manual snow course surveys and snow pillows all over the place. Here is a list of 400 or so (some are historic and no longer sampled) snow courses in BC. Many of those get visited every two or four weeks from Jan-Feb through June each year.

I've done the surveys, and you need to measure both snow depth, and moisture content. The process of manual measurement hasn't changed in decades - you drop a metal tube into the ground, pull it up, dig out the soil, measure the weight of the snow that the tube collected, and the depth of the snow. Of the two numbers, the overall moisture content is of greater interest. I hardly even look at snow depth when trying to decide if the water systems I run are facing a drought - the moisture content compared to historic trends is what matters most.

Even then, snow depth is only a guide. If you get high evaporation rates during spring freshet, or lots of wind and moisture loss, what appears to be a healthy snowpack in April can turn into near-record low runoff by June. This year that is exactly what happened in my region. We had a good snowpack, with normal amounts of water equivalent in April, but by June, very little runoff to the reservoirs had taken place. This mostly affected the low and mid elevation watersheds in the Okanagan. The really high elevation watersheds such as Mission Creek had normal runoff, while adjacent watersheds such as Mill and Hydraulic Creeks ended up with varying levels of drought.

More data is always a good thing, but the moisture content matters more than the depth. And even if the data looks promising, that can change in a matter of weeks. You never really know for sure how much water you're going to get until the reservoirs stop filling.

Re:Snow compactness?

[PeterBrett]

Does it account for the fact that almost-melting snow will layer more compactly then sub-30C snow which is extremely crystalline and less likely to make compact layers?

Disclaimer: I haven't RTFPaper. I'll do that tomorrow, since it's of professional interest. What follows is guesswork.

Imagine you've got an RF antenna on a flat surface. Above the flat surface is a layer of snow of constant thickness h. Now place at least three point RF sources at different angles theta and distances r from the antenna. Assume that all the sources lie above the snow.

Now, the distance that each RF signal must travel through the snow, x, will depend on d and theta, but not on r. You'll find that x is linearly dependent on d. Let t = alpha * x be the increase in signal flight time due to the snow. alpha will be a function of the dielectric properties of the snow.

So, basically, I don't think you can tell the difference between an increase in d and an increase in alpha, and ergo, the answer to your question is, "No." On the other hand, IIRC there's actually very difference in terms of dielectric properties between different types of snow, so it could be that the effect is dominated by the snow depth.

Re:Snow compactness?

[PeterBrett]

Heh. Having read the paper, it seems that it's not that simple, unfortunately. However, they seem to have assumed a standard model for snow:

We assumed density of 240 kg m3 for the new snow and a snow temperature of 2C ( = 1.48 i2.76 × 104), after Jacobson [2008].