FRESNO -- Farmers and water officials throughout California wait each year for forecasts about snowmelt roaring down the Sierra's granite canyons -- precious water for the long summer.
But the forecasts are only estimates based on averages of past seasons, snow-sensor readings and monthly measurements from mountain meadows above big rivers. The forecast can be wrong, leaving farmers with too much or too little water in the growing season.
One big reason: Nobody measures snow around jagged ridges, plunging ravines and deep forests in the 400-mile-long Sierra. That's a huge swath of the high country where the size of the snowpack is unknown.
As the climate warms and snowfall dwindles in coming years, officials will need to measure more of the Sierra to improve runoff forecasts for farmers and the growing population, say University of California at Merced scientists.
The researchers think they've found a way to do it by expanding monitoring around the remote sensors, called snow pillows, which are mostly in flat meadows. Scientists propose to surround them with instruments to check snow levels daily all over the landscape.
Someday, there could be thousands of these instruments in the Sierra.
"We're hoping to design a new system of doing things up there," said engineering Professor Roger Bales.
As head of the Sierra Nevada Research Institute at UC Merced, Bales has worked on this idea for five years at Gin Flat in Yosemite National Park. He and fellow researcher Bob Rice use ultrasonic monitors, devices that emit sound waves, to track snow's depth. They also use satellite images to see how much of the area is covered with snow, so their estimates won't include dry spots.
In the next year, Bales and Rice will expand the project to the mountains around the American River, above Sacramento. The $200,000 expansion will be funded using state bond money.
Progress depends on funds, early success
How quickly the technology takes hold throughout the Sierra, and begins to influence the accuracy of forecasts, will depend on funding and the success of these early efforts. No timeline has been set.
The ultrasonic monitors are an established and affordable technology, researchers said. The cost is about $1,200 per monitor, which includes devices such as transmitter links to send data via satellite or cell phone.
Mounted on poles about 30 feet high, the monitors send sound waves to the snow, and the waves bounce back to the monitor.
As snow accumulates, waves travel a shorter distance and bounce back faster. Researchers can calculate the snow's depth if they know how fast waves return to the monitor.
Rice said tests at Gin Flat revealed some snow pillows may be overstating the snow amounts by 25 percent, mainly because they are purposely located to measure meadows that catch a lot of snow.
The snow pillows are 8-by-10-foot stainless steel tanks filled with antifreeze. They weigh snow and help state officials determine how much water is in the snowpack, which is information used in the calculations made by the UC Merced researchers.
But they are not accurate reflection of the entire area, Rice said.
"The snow stays the longest in these (meadows) when everything else has melted around it," he said.
Frank Gehrke, California's snow survey chief, said the sound sensors also might solve the problem of measuring snowfall at elevations of more than 10,000 feet, especially in the rugged southern Sierra.
In late May and early June, a lot of snow has melted at lower elevations. But high up, there often is snow that won't melt for weeks. Sometimes, runoff is much stronger in August because of the high-elevation snow.
"We're really unable to quantify the amount of snow up high, and it makes a difference," Gehrke said.
State officials add that their runoff forecasting relies heavily on mathematical models that don't work well if a winter is much drier than average. If the climate warms, there may be many dry winters, scientists say.