From afar, the rows of knobby grapevines blend into the landscape of pink-blossomed almond trees and fragrant citrus. But get up close and you’ll see something strange: The trunks of the vines are standing in several inches of glistening, precious water.

These grapes, at the Kearney Agricultural Research Center in California’s San Joaquin Valley, are part of a grand experiment that many hope will help solve the state’s deepening water crisis. Here, in the state that provides some 40 percent of all the fresh produce grown in the United States, a 20-year-long drought has left growers and communities desperately short of water. To make up the persistent shortfall from rain and snow, they are pumping groundwater—and doing so far faster than water can trickle down from the surface to replenish underground aquifers.

The drought has only amplified an old problem: Californians have been overusing groundwater for a century, in part because it was unregulated. That changed in 2014 with the passage of a landmark state law requiring local water agencies to control the overdraft by 2040. They’re now scrounging for solutions.

One popular idea, which scientists are testing on the flooded grapes at Kearney, is to “recharge” overdrawn aquifers with water that would otherwise flow to the sea unused—the torrential precipitation that often comes during winter, for example, when farmers don’t need it. Farmers and water managers alike are trying to capture some of those precious pulses by intentionally flooding farm fields, wetlands, and other key places. Letting that water recharge California’s in-the-red aquifers would be a cheaper, more ecologically sensitive, and effective way to prepare for drought, proponents argue, than building more dams and reservoirs.

“It’s obvious we’ve been living well beyond our means,” says Aaron Fukuda, the manager of the Tulare Irrigation District in the San Joaquin Valley, one of the most seriously overdrawn regions of the state. “We all know we’ve either got to shrink the demand or increase the supply… and groundwater recharge is becoming the go-to solution.”

Water onto wine

Don Cameron has been thinking for decades about how to get water back underground.

Cameron has managed Terranova Ranch, a 5,500-acre farm in the San Joaquin Valley, since 1981, and he has long relied on groundwater. In the early days, that was easy: He and his team would hook up some 100-horsepower engines to wells and pump away. But over time, some wells dried up. Others had to be punched deeper. The ranch had to start running 200-horsepower engines just to keep the water flowing.

It was obvious to Cameron the problem would only worsen. He worried it would put him or his neighbors out of business; water is one of farming’s non-negotiable ingredients.

A solution also seemed obvious. If the issue was too much water coming out of the ground, wouldn’t it be simplest to try to put some back whenever an opportunity arose?

That opportunity appeared in February of 2011. After a dry year, an unexpected storm dumped snow and rain, sending a long glug of extra water down the King’s River, which snakes along the edge of Terranova Ranch. Cameron diverted some of it into a series of deep canals and ditches built in anticipation of this exact moment, set up a series of small pumps, and pumped enough water onto a field of wine grapes to flood them nearly two feet deep.

His neighbors were aghast, thinking there was no way the plants would survive a drowning. But Cameron’s intuition told him otherwise—the wild ancestors of grapes evolved along rivers, after all. “I was pretty damn sure they could handle having their feet wet,” he laughs.

The water kept disappearing into the ground, so he just kept pumping—all the way through June, well after buds burst and leaves unfurled, for a total of 13 feet of water over the season: enough to cover a football field to well above the field-goal crossbar.

The grapes survived just fine. The bigger value of the experiment, though, was to pique others’ interest. Cameron talked about his experiment constantly, hosting taco-truck talks at the farm and evangelizing the idea to scientists, policymakers, and farmers alike.

The following year, California sank into an epically bad phase of its long-running drought, one that would deepen and linger until 2016. So little rain and snow fell that farmers in some parts of the San Joaquin Valley had their surface water deliveries cut to as little as 40 percent of normal. Groundwater use, which makes up about 40 percent of all water use even in a wet year, shot even higher.

Like Cameron, other farmers found their wells running dry. Entire communities started to hear their pumps sputter. In some places the ground has sank more than 20 feet because its underbelly was drained of water.

That was the context in which lawmakers managed to push through the Sustainable Groundwater Management Act, or SGMA (pronounced “sigma”) in 2014. California became the last state in the West to put limits on groundwater use.

“It took over 100 years for our groundwater oversight to catch up with our surface water, but we got there,” says Kaymar Guivetchi, an engineer at the California Department of Water Resources.

Refill the bucket

There are really only two ways to stop emptying a groundwater bucket, explains Alvar Escriva-Bou of the Public Policy Institute of California (PPIC): draw less water, or fill the bucket back up.

“It’s the simplest math in the world,” says Rosemary Knight, a geologist at Stanford University.

Using less water would probably require fallowing farmland, causing losses for growers and their communities. Compared to that, refilling the bucket is much more palatable—and so Don Cameron’s ideas about recharging groundwater got very, very popular. 

“It was like someone flipped the switch,” says Cameron. “They saw that this was going to be a thing that was sustainable and could keep them in farming. It sank in all of a sudden.”

The idea is straightforward: When extra water rushes through the state’s waterways, some of that could be “peeled off” and diverted to fields, recharge basins, wells, or other floodable sites.

In some ways, this would mimic California’s historical, pre-settlement water cycle, says Guivetchi. Before the era of dams and long-distance canals, California valleys flooded much more frequently, when mountain snowpack melted in huge spring gushes or big storms dumped heavy rains. The floodwaters would seep down into aquifers below. But enormous 20th-century state and federal water projects replumbed the region, draining natural flood zones, channelizing rivers, and fundamentally disrupting the natural water cycle to support agriculture and hydropower.

“The water cycle wants to operate the way it’s done for millennia. But our institutions have sliced and diced it to the point where it’s become dysfunctional,” Guivetchi says.

Restoring more of the natural flood cycle to refill aquifers is something that many groups in the state agree on. The challenge, says Guivetchi, is how to get it done, not on a few hundred acres, as at Terranova Ranch, but likely on many thousands.

Putting water back on the land

The groundwater in the San Joaquin Valley is overdrawn by about 2 million acre-feet a year— about one-third as much as all Californians use in their homes. (An acre-foot would cover one acre, a football field, in one foot of water.)

No one yet knows exactly how much of that deepening debt can even be repaid. Local water districts’ SGMA plans currently propose to fill about 40 percent of the hole with recharge. But an independent analysis from the PPIC found that recharge could fill only about 25 percent of the overdraft.

Ideally, water would be injected in confined zones where deposits of coarse gravel or sand offer a direct channel to an aquifer. Such “sweet” zones are rare, but Fukuda’s Tulare Irrigation District already uses one, and the state is spending $12 million to find more.  

Stanford researchers led by geophysicist Knight have figured out how to find them from the air or by driving overland on a 4x4. Equipped with an instrument that maps subsurface geological structures by measuring their electrical conductivity, the team can tell gravel and sand—through which water travels easily—from clay, which stops its downward motion. 

Hyper-effective recharge zones probably won’t be enough, though. Farmers will also need to be willing and ready to flood their fields.

“This is why we’re here at the research station,” says hydrogeologist Helen Dahlke, as she pulls her boot out of squelching mud in the flooded grapevines at Kearney. “Torturing grapes!” interjects a laughing Shulamit Shroder, a graduate student.

Many farmers Dahlke works with want be part of the solution. But they need to know that their crops will survive the flooding, and how to keep pollutants like nitrate—a common fertilizer component and a major contaminant in local waters—from leaching to the aquifers. That’s what Dahlke’s team is testing here. They flooded this field two days ago; now, they’re watching whether and how nitrate moves down through the topsoil.

She leans over and switches on a small portable vacuum pump, gently pulling a sample from a buried tube, her face reflected in the water below. “There are definitely a lot of potential tradeoffs in agricultural recharge,” she says, which is “why need to figure out how to use the sweet spots.”

Finding the extra water

A bigger problem than finding the sweet spots is finding the water itself.  But there is potentially available water, even in the driest years—usually during intense winter storms.

Climate change is intensifying those downpours even as dry spells get drier. It’s also changing snowmelt patterns so spring floods come faster and bigger. The Department of Water Resources recently projected that by the 2070s, flood events on the Merced River—at the north end of the San Joaquin Valley—could be 600 percent bigger than today’s highest flows.

If floodwaters could be peeled off and directed to groundwater recharge, the risk of damaging floods could be decreased at the same time, says DWR engineer Ajay Goyal. But there are legal and logistical hurdles. Complicated fights about who has rights to the “extra” water—and what exactly constitutes “extra”—are already brewing. Miles of aging, expensive infrastructure would need to be updated to accommodate even larger water pulses.

Cameron was years ahead. In the 1990s, preparing for his flood experiment, he built a big ditch at the edge of the property to hold the extra water. (“We didn’t get permits or anything, and I should probably be in jail, but we just had to try it,” he says.) After his success in 2011, he decided to build an even bigger system—in coordination this time with county and state governments and his neighbors and sized for a further climate-changed world.

On an early spring day, Cameron stands on concrete headwater gates at the southwest corner of the ranch. Squinting in the sun, he gestures to his left at a tiny trickle that winds across the wide, tussocky plain. “That’s the North Fork of the Kings River,” he says. “Impressive, isn’t it?”

But during wet times, he explains, the plain roars with water—and that’s when his canal will shine. He gestures to the right. A banked-earth canal, 15 feet deep and 70 across, stretches a mile along the edge of his property. Then it splits; one leg heads north for another mile. The other branches east and is slated to run for 12 miles and serve many of his downstream neighbors.

It’s almost comically enormous compared to the meager dribble in the riverbed. But it’s what they’ll need when the water comes, says Cameron. With this system and the right weather, “I could just about put as much water back down as we use in a year,” he says.

For now, it’s a waiting game. Last December was one of the wettest on record in California, and Cameron got his hopes up. But in the kind of weather whiplash that has become all too familiar, the rest of the winter was almost record-breakingly dry. His canals will have to wait another year, or two, or 10.

“We’re waiting. But we’ll be ready,” he says.