Snowpack Percent of Normal, State by State

It is the water held in the snow right now, divided by the median amount of water held in the snow on that same calendar day across a fixed thirty year period, and it is only meaningful if you know which thirty years. Every figure on this site uses the 1991 to 2020 period, which is the one the agency currently publishes. The same station on the same day came out at 89 percent against 1991 to 2020 and 94 percent against 1981 to 2010, so a percentage with no period attached is not a fact.

Because depth does not tell you how much water is there. A foot of cold dry powder and a foot of settled maritime snow can differ by a factor of three or four in the water they contain. Snow water equivalent is measured by weighing the snow: the instrument is a fluid filled pillow lying on the ground under the pack, and the pressure in it reports the weight of everything above. That weight is the number a river cares about, which is why the water supply network measures it and why every figure here is in inches of water rather than inches of snow.

12, because 12 are all that have a network to read. The federal station inventory lists no active snow telemetry stations in thirty-seven states, and two in South Dakota, both in the Black Hills, which is not enough to describe a state. The network exists to forecast summer water supply in the West, where a year of water arrives as winter snow. East of the Rockies rain falls through the growing season and no water supply forecast ever needed a mountain snow network, so nobody built one.

The measurements are the official ones. They come from the United States Department of Agriculture's Natural Resources Conservation Service, which operates the snow telemetry network, and they are federal public domain records. Our per station values and percentages reproduce the agency's own published figures exactly: all 118 Colorado stations matched on value, on median and on percentage when checked against the agency's own report for March 15, 2026. The statewide and basin totals are OURS, not the agency's, because the agency publishes no statewide or basin index through any interface a site can read. Every place one appears on this site says so.

It depends enormously on where, and the spread across these 12 states is the reason this question has no single answer. Taking each state's stations and finding the middle of their own median peak dates, Arizona peaks around March 2 and Montana not until April 20. That is roughly seven weeks of difference driven by latitude and by how long each region stays cold enough to keep adding rather than start losing. April 1 is the traditional benchmark date for the whole West, and it is genuinely close to peak in the northern Rockies and well past it in the Southwest.

More than most people expect, and the range between states is enormous. Taking the median across each state's stations of that station's own median peak water content, Washington tops the list at 32 inches of water and Arizona sits at 6.7. The single deepest station in Washington typically peaks at 103.2 inches of water, which at a typical settled density is well over twenty feet of snow standing on the ground.

It is an input to both questions and not an answer to either. How much of a snowpack reaches a river depends on spring temperatures, on rain falling on snow, on dust darkening the surface, and above all on how dry the soil under the snow was going into winter, because dry soil absorbs melt before any of it runs off. None of that is measured by a snow pillow. The agencies that do make those forecasts are the Natural Resources Conservation Service for water supply and the National Weather Service river forecast centres for river flow, and both are linked from every page here.

Because some of them are broken, and this is the single largest data quality problem in the network. A station whose sensor has failed reports zero, and the agency marks that value as estimated rather than valid, but the daily files this site reads carry no such marks. So any station reporting exactly zero water against a normal of an inch or more is held out of our totals. Measured against the agency's own quality flags for March 15, 2026, that rule held out 91 stations across these 12 states and the agency had independently flagged 72 of them, or 79 percent, as something other than valid. The rule errs in one direction: in a genuinely early melt year a station that really has melted out gets held out too, which biases the figure very slightly high. A failed sensor counted as zero would be the larger error.

It is yesterday’s, which is both the freshest and the most correct choice. The agency regenerates its daily files overnight with about a day of lag, so today’s file still holds last year’s data. It also publishes every figure as a start of day value, and the start of a day is the same instant as the end of the day before, so reading yesterday’s file and dating it today reproduces the agency’s own published number exactly. If the file does not reach the current year, the page marks the reading as stale.

Not as a data product. The agency publishes its basin numbers in monthly water supply outlook reports, as PDFs, and its interactive map computes basin figures in the reader’s own browser from the station values. The basin figures on this site are ours, computed by summing station values and dividing by the sum of the same stations’ medians, and they are labelled as ours everywhere they appear. They will not always match a number from an agency report, and where they do not, the agency’s is the one to use.