SAN JOSE, Calif. — It was the snowpack reading that spawned a hundred headlines. “California ties 1952 record for all-time Sierra snowpack,” proclaimed KTVU. “California’s snowpack soars to record high after 17 atmospheric rivers,” trumpeted the Washington Post.
State officials largely seemed to agree. “As of right now, it’s looking like this year’s statewide snowpack will probably, most likely, be either the first — or second — biggest snowpack on record dating back to 1950,” Sean de Guzman, manager of the California Department of Water Resources monthly snow survey, declared the day of the official April reading earlier this month.
But this year wasn’t a record at all. It only appeared that way in large part because of the state’s shifting definition of a “normal” snowpack, which critics say obscures the true impact of climate change.
And, in a closer look at the state’s seven decades of snowpack data, 1952 — the year atop the state’s data — wasn’t a record either. That top honor should go to 1983, which cinched first at 231% of normal, a new analysis by the Bay Area News Group found, when “normal” is considered the average dating back to 1950, the start of the state’s recordkeeping.
Over that period, 1952 is tied for second with 1969 at 226% of normal. This year actually came in fourth, at 222% of normal.
You won’t find those percentages in the California Department of Water Resources’ official records. But in an interview this past week, the agency didn’t dispute the data analysis behind the reordering that gives 1983 its due as the biggest, baddest Sierra snow year on record.
What explains this history redefining shift? Twin issues in how the state compiles the statewide snowpack figures are to blame.
So we ran the numbers, compiling 74 years of April snowpack data from over 300 locations called “snow courses” scattered throughout the Sierra, solving both issues plaguing the state data. The changes are not massive. But even small shifts can have an impact on scientific models. And they can change which year breaks a record.
The state’s data misstep
The first issue with state’s official records was a more straightforward record-keeping misstep, which inflated the state’s snowpack data for 1952 — the supposed record year.
There are nine low-elevation snow courses, which show up in the state’s reports that are not supposed to have an effect on the statewide averages. They are excluded because they get very little snow in a typical year. So when they do get even a small amount of snow, their “percent of normal” calculation gets turbocharged.
But here’s the thing — the stations only started to get excluded from the statewide statistics gradually, starting in the late 1990s. They were included in historical data before then, hyperinflating some years’ statewide “normal” totals.
The problem is particularly apparent in 1952, the supposed “record” snowpack year. That year, McElroy Pass and Burney Springs, both of which are excluded from later reports, reported 1200% of normal snowpack and 1020% of normal snowpack respectively. Because those stations got fed into the statewide average that year, they inflated it to first place. In our analysis, we excluded all of these faulty snow courses.
“This is maybe where the DWR (state water agency) mission departs a bit from the sort of — you know — best scientific practices that one might want to do, like, in a scientific study, where that (issue) would probably be flagged by a reviewer,” said John Abatzoglou, professor of climatology at the University of California, Merced. “I think they’re more sort of focused on what’s happening right now.”
What is normal?
There’s another issue with the state’s hyper-focus on the here and now. All of the statewide snowpack statistics are reported based on the “percent of normal” snowpack. But state water managers have changed the definition of what’s considered “normal” five times since 1950. “Normal” snowpack back in 1952 was calculated after-the-fact based on data from 1946 to 1995. Normal in 2023, meanwhile, was calculated based on data from 1991 to 2020, a period marked by many years of drought.
The reason they change the definition of normal is “to keep pace with climate change,” according to de Guzman. Using these shifting definitions of normal improves snowmelt flood forecasts, meaning that the state is better able to predict when snowmelt could turn into a disastrous problem downstream — a major concern in the coming weeks, especially in the Southern Sierra.
But there is an ironic consequence to shifting the goal post — by changing the definition of “normal” to keep pace with climate change, experts say the state masks its impact in the historical data.
“We’re sort of looking in the rearview mirror, but we’re not looking in the rearview mirror on the last 100 years… we’re looking in the rearview mirror in terms of our own memories,” Abatzoglou said. “And, you know, we’ve been on this bumpy, bumpy ride with climate change — uphill towards warmer climates.”
The closer you get to the present, the more the state’s definition of “normal” includes a bunch of unusually warm years with less-than-normal snowpack. That artificially inflates the most recent percent of normal numbers in comparison with earlier eras, like the 1950s and 1960s.
To get around the shifting definitions of normal, the Bay Area News Group collected all of the state’s snowpack data and applied a single definition of normal, which stretched the entire record period from 1950 to 2023. By putting all the data on the same rate, experts say you can much more readily see the impact of climate change.
More evidence that California snowpack has declined over time can be found in what experts call the “snow line,” which essentially marks the elevation at which snow begins to accumulate. The snow-water line has risen by about 500 feet due to climate change.
“We are going to see the lower stations start to get less snowfall,” said Andrew Schwartz, lead scientist at the Central Sierra Snow Lab, a research station maintained by UC Berkeley located at Donner Pass. “We’re also seeing, at least at the Central Sierra Snow Lab, a shortening of the snow season. … And that’s a trend that’s been noticed in a lot of other regions as well.”
The main thing that scientists look at to gauge how much snowpack we have is something called “snow water equivalent” or “snow water content.” That measure is the amount, in inches, of water that would be produced if all the snowpack melted at once. From 1950-1986, the Sierra had 29.2 inches of snow water content. That dipped to 25.6 inches from 1987 to 2023, according to a Bay Area News Group analysis. The EPA, meanwhile, documented a 23% decline in snowpack levels at stations across the Western United States from 1955 to 2022.
Since California gets nearly a third of its drinking water from snowmelt, those declines are a huge deal. In that same time period, California’s population nearly quadrupled to 40 million, putting even more strain on the water supply.
So what?
State water officials point out there will always be issues around collecting statewide snowpack data. It’s hard to get to some of the snowiest courses in the winter. And some snow stations are decommissioned over time.
“A lot of folks — especially academics and researchers — have asked us to try to redo this (statewide snowpack data) to put everything on the same playing field, like… what you’re doing,” de Guzman said. But he explained that there are technical limitations on their ability to rerun the numbers. And state officials don’t really go back to adjust numbers that were previously published, especially since those numbers were used in forecasts at the time and had real world impact.
A change to one data point or even one entire dataset is not going to radically change our understanding of our warming climate, experts say. But having a clear record of how our state’s climate is changing over the long term is incredibly important for understanding where we’re headed as our planet warms.
“No one data point is ever going to prove or disprove climate change.” Schwartz said. “It’s about the longer trends, and the comparison years are important. But when we look at the overall, you know, over the entire record, that’s when we really start to see this (climate change) trend emerge, and when we stop getting lost in the weeds.”
