Why California Drains Its Reservoirs Before a Storm Even Arrives

Fifty-eight feet. That's how far Lake Oroville's surface climbed in the last three weeks of December alone, after a run of atmospheric river storms hit Northern California. By New Year's Day the reservoir sat at 70% of capacity — well above its historical average for the date — and the California Department of Water Resources was already doing something that looks, on the surface, counterintuitive: dumping more water out.

DWR pushed releases to the Feather River up to 15,000 cubic feet per second on Jan. 2, 2026, with plans to go to 25,000 cfs within days, and said it would open Oroville Dam's main spillway on Jan. 5 as inflow was projected to hit 50,000 to 70,000 cfs. None of that was a response to a failure. It's the routine, and it's required: between mid-September and May, federal Water Control Manual guidelines set by the U.S. Army Corps of Engineers force DWR to hold a reserved block of empty space in Lake Oroville — the "flood pool" — at all times, specifically so the reservoir has somewhere to put the next storm's runoff. A full reservoir looks reassuring. To a flood engineer, it's a reservoir with nowhere left to send an incoming surge except downstream, uncontrolled.

The logic scales down to any dam with a flood-control mission: release water ahead of the storm, on a schedule you control, to a river channel that can absorb it, rather than get forced into an emergency release once the reservoir is already full and the storm has arrived. It's the difference between draining a bathtub a little at a time and watching it overflow onto the floor.

What changed by 2026 is how early operators can start. A March 2025 assessment from the Center for Western Weather and Water Extremes at UC San Diego's Scripps Institution, done with DWR, Yuba Water Agency and the Corps' engineering research arm, found that better atmospheric-river forecasting lets operators time those releases with far more precision — a strategy called Forecast-Informed Reservoir Operations, or FIRO. Instead of holding a fixed, conservative amount of empty space all winter just in case, operators can watch the storm forecasts and adjust in real time, releasing early when a big system is coming and holding more water back when it isn't. DWR's state climatologist, Michael Anderson, called reservoirs "critical" to safeguarding communities as storms grow more extreme and droughts grow more severe, and said the Yuba-Feather assessment was "an important step toward improving flood protection ... while managing California's valuable water resources for decades to come."

The payoff isn't only fewer floods. DWR Director Karla Nemeth described FIRO as "a critical step towards making more floodwaters available for groundwater recharge when we see extremely wet storms," since a reservoir that can time its releases more precisely doesn't have to dump water as conservatively as one working off fixed calendar rules. Yuba Water's director of resource planning, John James, put a number on what that flexibility is worth: paired with a second spillway now in design at the agency's New Bullards Bar Dam, FIRO gets the region "essentially gaining the amount of flood storage that would historically only be created through building additional infrastructure," without pouring new concrete for a bigger reservoir.

The spillway is the physical half of that upgrade. New Bullards Bar's planned Atmospheric River Control Spillway would sit with its gates 31.5 feet lower than the dam's existing spillway, letting the agency release up to 35,000 cfs before a threatening storm arrives — while there's still enough room downstream to carry the flow safely — rather than waiting until the reservoir has no choice. The project is fully designed and Yuba Water is chasing state and federal grants, with construction targeted for as early as 2027.

None of this is unique to the Feather and Yuba watersheds. FIRO already governs releases at Lake Mendocino on the Russian River and is being phased in at Prado Dam on the Santa Ana River in Southern California — a quiet, forecast-driven reshaping of how the state's reservoirs are run that gets far less attention than the storms themselves. Flood risk was never just about a storm's strength in the first place, and it isn't only about the tides that can flood a street with no storm at all; upstream, it's about how much room is left in a reservoir to catch what falls. That's the number DWR is managing every time it opens a gate on a clear day before the rain even starts.