upslope
How the models work, why snow is harder than rain, and how we grade ourselves.
Every number on this site comes from public data produced by the National Weather Service. Nothing here is proprietary — what upslope adds is putting all 32 New England mountains side by side, showing the honest range instead of one hyped number, and grading itself in public afterward.
How a snow forecast is actually made
Weather models take today's observations and simulate the atmosphere forward. They predict liquid precipitation first; snow depth comes from multiplying that liquid by a snow-to-liquid ratio that depends mostly on temperature. That two-step process is why snow forecasts are harder than rain forecasts: an error in either the moisture or the temperature compounds into the number you actually care about. A fifty-mile shift in storm track or a two-degree temperature difference can double or halve a mountain's total.
Snow-to-liquid ratio, and why powder days feel different
One inch of liquid water does not produce a fixed amount of snow. In cold, dry air it can fluff up past 15:1 — fifteen inches of light powder. In a warm storm near freezing it falls closer to 8:1: dense, heavy, and hard to ski. Two mountains reporting "ten inches" can deliver completely different days. Where the data supports it, upslope shows the ratio alongside the total, because the number alone doesn't tell you what you'll be skiing.
Why we show a range instead of one number
The National Blend of Models runs roughly thirty forecasts side by side. When they agree, the range is narrow and you can trust the number. When they disagree, the range is wide — and a site showing you a single confident total in that situation is hiding the most important thing about the forecast. We show the 10th to 90th percentile: one run in ten came in below the low end, one in ten above the high end. A wide spread is information, not a failure.
How we grade ourselves
Every forecast upslope publishes is written to a database before the weather happens. The next day, volunteer observers from the CoCoRaHS network — people with standardized gauges in their backyards — report what actually fell. We take the median of the gauges near each mountain and compare. The results, including the misses, are on the scoreboard and in the per-storm receipts. Nobody else in New England publishes this, which is either an opportunity or a warning, depending on how you look at it.
The acronyms
NWS — the National Weather Service, taxpayer-funded and public, the source of nearly everything here. NDFD — the forecaster-adjusted grids behind our 7-day numbers, hand-tuned by meteorologists at the Burlington, Gray, Boston, and Albany offices. NBM — the National Blend of Models, roughly thirty forecasts run together, where our ranges and odds come from. HRRR — a 3km model refreshed hourly that resolves the rain/snow line, powering the hourly strips. AFD — the Area Forecast Discussion, the forecasters explaining their own reasoning in their own words. CoCoRaHS — the volunteer observer network that referees our scoreboard. Full definitions on the about page.