Getting power out there might be as simple as laying a UL rated outdoor extention cord down when the ground freezes and picking it up when snow melt happens in the spring, since burying a 120v cable deep enough to feel safe would be a hurculean task.
The way that I look at it, millions of homes hang holiday lighting outside with extension cords every year. If their inflatable snowman doesn't have a problem with it in the snow, why should a solid state meter that could probably run on a car battery for a month...if not the entire season?
I have a pocket 16,000 mAh pocket cell phone emergency battery pack that will power the NodeMCU that I assembled for solar sensor collection running for 4-5 days between recharges, and I made no effort to trim power use in software. It sends 5-second updates via wifi, and leaves the radio on between transmissions.
My idea for the snow sensor was going to depart with the common homebrew setup with a fixed-mount sensor measuring a single spot. I was thinking of using a round patio table (with the appropriate surface prep) as the snow board, placed in the middle of the yard. There will be a vertical pole inserted into the umbrella hole, with a stepper motor to rotate it. The sensor electronics will be at the top of the mast, on an offset arm. Measurements will be taken by rotating the mast/sensor to all 8 cardinal/ordinal directions, and then averaging the measurement. Seems more in line with the suggested NWS manual measurement process, and there's less wondering if the sensor is blocking totally vertical snowfall.
I plan to put an optical sensor in the table surface at due north, which will be used to calibrate "home" on the stepper motor as well as detect the onset of snowfall via the loss of the laser signal. There is probably no benefit to turning the mast if it's not snowing, to save both mechanical wear and power of running from battery/solar.
Comments on the methodology or idea?????
-Arthur