Groundwater Recharge in California

California agriculture runs on two supplies that trade places. In wet years, surface water from rivers, reservoirs, and the state and federal projects carries most of the load and aquifers rest or refill. In dry years the aquifer becomes the supply of last resort.

Statewide, groundwater provided about 30 percent of agricultural water in wet 2017 and 2019, 63 percent in the 2015 drought, and 57 percent in 2022. In the Tulare Lake region the swing runs from 22 percent in wet 2017 to 90 percent in 2015, when surface deliveries collapsed from over 9 million to about 1.1 million acre-feet.

Subbasin reporting under SGMA shows the same mechanism at field scale. Across the subbasins that file annual reports, groundwater's share of agricultural water use ran near 67 percent through the critical years of 2021 and 2022 and fell to 36 percent in wet 2023. The swing is sharper in the southern San Joaquin Valley: in the Kings subbasin the share moved from 90 percent in critical 2021 to 38 percent in wet 2023, and in Kern County from 73 percent to 27 percent. When surface water arrives, pumps rest and aquifers recover. Every acre-foot of reliable surface supply is an acre-foot the aquifer keeps, and in wet years it becomes water the aquifer can take in.

Decades of dry-year reliance have accumulated into overdraft. For the San Joaquin Valley, the Unified Water Plan projects a gap of about 2.4 million acre-feet a year by 2040, the year SGMA requires critically overdrafted basins to reach balance. Recharge is how wet years pay that balance down.

When water tables fall far enough, fine-grained layers in the aquifer system compress. Some of that compression is elastic and recovers when levels rise. Past a threshold it is inelastic: the clays collapse, the land surface sinks, and the storage space is gone for good. DWR's InSAR monitoring shows nearly 8 feet of cumulative sinking since June 2015 in the most affected parts of the San Joaquin Valley. The damage runs through infrastructure as well as storage: subsidence has cut the Friant-Kern Canal's capacity through its middle reach by more than half, and canal repair and recharge are two halves of the same fix, since raising groundwater levels is what slows the sinking.

A study published in PNAS in July 2026 extended this picture to the Sacramento Valley, which had largely been considered the state's stable aquifer. Combining satellite radar, GPS, well records, and satellite gravity data for 2016 to 2022, researchers at UCLA and Caltech found that large areas shifted abruptly to mostly irreversible compaction in 2021, during the 2020 to 2022 drought. Ground in the most affected areas sank as much as 20 inches per year, against a reversible response of well under an inch, and permanent storage loss rose roughly fivefold, to about 160,000 acre-feet per year in 2021 and 2022.

The finding that matters most for management: recovering water levels do not mean a recovered aquifer. In many areas the ground did not rebound after seasonal recharge, confirming that part of the storage space itself was lost. Recharge deferred can mean capacity that no longer exists when the next wet year arrives. That is the argument, written in the land surface, for moving water underground in the years when it is available.

What sets the cost. Soils in the Kings study sustained infiltration of about 2.5 inches per day after an initial rate near 8 inches per day, through flooding events lasting up to 20 days. At that sustained rate, roughly 10 acres of ground absorbs 1 cubic foot per second of delivered flow. The binding constraint in that project was not the ground. It was the pipe: diversion and pumping capacity limited application to about 0.26 inches per day, an order of magnitude below what the soil could take. Conveyance and turnout capacity, not infiltration, usually decide how much water a winter can put underground.

What the crop can take. Roughly half of San Joaquin Valley cropland is considered suitable for on-farm recharge. Field trials give the tolerance picture directly. Almond orchards on moderately drained to well-drained soils received about 24 inches of applied water in late December and January in each of two years with no adverse effect on tree water status, canopy development, or yield, and more than 80 percent of the applied water percolated past the root zone. Alfalfa tolerates heavy winter flooding with minimal yield loss, and researchers estimated that flooding all suitable alfalfa acreage with 6 feet of winter water could bank on the order of 1.6 million acre-feet a year. Wine grapes showed no crop damage across two seasons of spring flooding for direct recharge, and pistachios showed no significant yield penalty. On annual crops such as tomatoes, recharge is run during the fallow window.

Who pays and who is paid. Districts have built incentive structures rather than mandates. Westlands and Arvin-Edison have paid landowners $30 to $100 per acre-foot to take surplus district water and recharge it on their land. Landowners who conduct recharge commonly receive 50 to 90 percent of the diverted volume back as a credit against future pumping, with the remainder left behind for the basin. Pajaro Valley rebates half its groundwater pumping fee, currently $323 to $452 per acre-foot, for water recharged. Growers do carry real costs on their side of the ledger, including meters, dual-use irrigation, checks and turnouts, and the labor to run and monitor a flooded field through the winter.

What it is worth. Valley growers have generally not been willing to pay more than $300 to $500 per acre-foot for new long-term supply, which sets the practical ceiling any recharge project has to price under. The return side is larger than the water bill. Analysis of the Valley's path to groundwater balance found that ending overdraft with limited flexibility means fallowing about 750,000 acres, with annual regional economic output down about $2.1 billion and roughly 21,000 jobs lost. Adding supply and allowing water to move to its highest use brings that to about 535,000 acres, $1.3 billion, and 13,000 jobs, a reduction of roughly 40 percent in the employment effect. Recharge is one of the supply-side levers in that calculation, and it is among the cheapest of them.

More than $520 million has been disbursed to local agencies since 2015 for sustainable groundwater planning and implementation, which is why cost-share, not full local capital, is the usual financing shape for a first project.

The Sustainable Groundwater Management Act of 2014 assigned groundwater to local control with a state backstop. Bulletin 118 defines the basins and subbasins; Groundwater Sustainability Agencies, formed by counties, cities, and water districts, adopt and implement Groundwater Sustainability Plans; and critically overdrafted basins must reach sustainability by 2040, avoiding six undesirable results that include chronic level declines, storage reduction, and land subsidence. SGMA governs demand and accounting. It does not supply water, which is why recharge capacity decides how much of the balancing arrives through wet-year supply rather than through reduced pumping.

The state's direction of travel is toward making recharge easier and faster. A Vision for the San Joaquin Valley, released by DWR in May 2026 and open for public comment through July 21, 2026, prioritizes near-term high-flow capture pilots, raising groundwater levels to slow subsidence, regulatory reforms that streamline recharge permitting, and a public dashboard for identifying recharge opportunities, with long-term investment in storage, conveyance repair, and subsidence remediation. It builds on the State Water Project Adaptation Strategy, the San Joaquin Valley Conveyance Study, and the San Joaquin Basin Watershed Studies, and feeds the California Water Plan 2028 under SB 72. The final version has not yet been published.

Legislation is moving in the same direction. AB 2026 would write the 90/20 method into the Water Code, add flood-recharge and location-specific diversion criteria as expedited routes through a water availability analysis, and extend the simplified minor-application path to private parties operating under an agreement with a Groundwater Sustainability Agency. It passed the Assembly 68 to 0 on May 27, 2026, cleared its Senate policy committees without a no vote, and was placed on the Senate Appropriations suspense file on August 3, 2026. Until it passes, the 90/20 method remains State Water Board practice rather than statute, and the fee timing in the table above stands as written.

The through-line across all of it: reliable surface supply and recharge capacity are what let basins reach balance with their agricultural economies intact.