SnowBase Data Methodology
A clear, consumer-accessible explanation of how physical snow sensors operate, how telemetry is normalized, how confidence is evaluated, and how condition scores are computed.
1. How Mountain Snow Depth is Measured
Measuring snow in remote, sub-zero mountain environments requires specialized automated instruments. SnowBase primarily ingests data from two federal sensor networks:
A. Ultrasonic Acoustic Depth Sensors (SNOTEL)
Mounted on an arm 15 to 20 feet above ground level, the ultrasonic transducer emits high-frequency sound pulses toward the ground. By measuring the round-trip speed of sound (adjusted for air temperature), the sensor determines distance to the snow surface, calculating total settled snow depth (SNWD) to within 0.5 inches.
B. Pressure Snow Pillows (Snow Water Equivalent - SWE)
A stainless steel or hypalon bladder filled with an environmentally safe antifreeze solution rests on the ground. As snow accumulates above it, the weight of the snowpack exerts hydrostatic pressure on a precision transducer. This directly measures Snow Water Equivalent (SWE)—the actual liquid water mass contained in the snowpack.
C. Climatological Observer Records (NOAA GHCN)
In Midwestern and Eastern trail networks (Michigan UP, Wisconsin, Tug Hill NY), manual daily measurements by cooperative observers and national weather stations record 24-hour snowfall (SNOW) and total ground depth (SNWD) at fixed morning observation times.
2. "Measured" vs. "Modeled" Data
We maintain a strict visual and algorithmic distinction between empirical observations and numerical model estimates:
- Measured: Direct, verified physical readings from an active SNOTEL or GHCN sensor.
- Modeled: Atmospheric simulations (High-Resolution Rapid Refresh - HRRR or GFS) utilized when a physical sensor is unavailable, frozen, or located between monitored mountain passes.
3. Data Freshness & Confidence Rating
Every location and station on SnowBase computes an independent Data Confidence Rating:
- High Confidence (80–100%): Telemetry verified within the past 6 hours from an active physical sensor with a synchronized 7-day forecast.
- Moderate Confidence (50–79%): Observation is between 12 and 36 hours old, or minor telemetry inputs are interpolated.
- Low Confidence (< 50%): Older observation (> 48 hours) or purely modeled estimate.
4. 30-Year Historical Climate Baselines
To compute % of Normal, SnowBase utilizes the official 30-year climate normals (1991–2020 baseline period) published by the USDA NRCS National Water and Climate Center. We use the median rather than mathematical average to prevent extreme anomalous winters from distorting typical snowpack expectations.
5. Condition Scoring Architecture
The 0.0 to 10.0 Condition Score is a deterministic function tuned for each of 5 sports disciplines:
Score = clamp(BaseDepthScore + 24hPowderBonus - TempThawPenalty - RainPenalty - WindPenalty, 0.0, 10.0)Scores are never presented as safety clearances. They reflect snowpack volume, surface softness, and weather comfort.