Forecast Informed Reservoir Operations

As weather extremes become the norm, USACE and partners pioneer a new way to manage water resources.

Although nobody knew it yet in January and February of 2020, the state of California was in the early months of what would become the driest three-year period in the state’s recorded history. But when rain fell in one of the region’s watersheds – the Russian River basin, which drains 1,485 square miles in Mendocino and Sonoma Counties – water managers at one of the system’s reservoirs, Lake Mendocino, were able to exercise a choice they wouldn’t have had under their existing management plan.

The 2019-2020 “water year” – which, in California, begins on Oct. 1 and ends on Sept. 30 of the following year – was when U.S. Army Corps of Engineers (USACE) and its partners at Lake Mendocino piloted a new approach, Forecast Informed Reservoir Operations (FIRO). Traditionally, decisions about whether to hold water behind a dam, or whether to release it downstream, are influenced or determined by conditions on the ground.

“Under our normal operating rules, all water within the flood control space would have had to be released,” said Patrick Sing, lead water manager for USACE’s San Francisco District. “But we had a temporary deviation agreement in place that allowed us to implement FIRO, and we had some dry weather forecasts in place at the time, so USACE was able to retain some of that water.”


In total, the amount of water held back was 20% more than it would have been under previous guidelines: an additional 11,650 acre-feet, enough to supply up to 24,000 households for a year. According to the 2020 Census, there were just over 34,000 households in Mendocino County at the time.

“One could argue,” he said, “that if we hadn’t been able to bank that water in early 2020, the lake would have gone completely dry.” – Nick Malasavage, Ph.D., PE, USACE

Household supply is an easy-to-understand measurement, said Nick Malasavage, Ph.D., PE, chief of Operations and Readiness for the San Francisco District – but it doesn’t give a complete picture of the value of the water resource, which provides water for agricultural, ecological, municipal, and industrial uses along a vast stretch of the Russian River. “That water supply burden is carried by Lake Mendocino from Ukiah to Healdsburg,” Malasavage said. “So the water that was saved in 2020 contributed to carrying that burden for multiple communities in at least two counties, not to mention the agricultural draw and the minimum flow requirements for fish species.” The Russian River and its tributaries provide habitat and spawning grounds for steelhead trout (a threatened species, under federal rules) and chinook and coho salmon (both endangered).

Lake Mendocino, where FIRO was first tested and implemented by the Corps of Engineers. Small and easy to manage and measure, Lake Mendocino was a perfect candidate as a pilot to demonstrate FIRO. USACE PHOTO BY BRANDON BEACH

That water came to feel more precious in the coming months as, according to Malasavage, Lake Mendocino likely equaled or surpassed its lowest level ever recorded, about 12,000 acre-feet, in the summer and fall of 2021. “One could argue,” he said, “that if we hadn’t been able to bank that water in early 2020, the lake would have gone completely dry.”

Weather Whiplash

The idea for forecast informed reservoir operations isn’t brand new. It’s been years in the making – and during those years, many Western states have cycled dramatically between extreme floods and drought. In California, rainy seasons are now ruled by atmospheric rivers – literally rivers in the sky, wide bands of moisture, often thousands of miles long, fueled by the evaporation of warm ocean waters and driven by strong winds – that release torrents of rain or snow when they make landfall. According to one of the Corps of Engineers’ partners in developing FIRO, the National Oceanic and Atmospheric Administration (NOAA), a strong atmospheric river can move up to 25 times the average flow of water at the mouth of the Mississippi River.

Atmospheric rivers are accounting for an increasing percentage of the West Coast’s annual precipitation and streamflow – about half, in California. The intensity of these events, which are often followed by long dry periods, has made water management difficult to achieve with the old playbook.


The good news is that while they’ve introduced unprecedented volatility to California weather, atmospheric rivers are increasingly easy to predict with today’s expertise and technology. The Center for Western Weather and Water Extremes (CW3E) at UC-San Diego’s Scripps Institution of Oceanography is helping move the leading edge with models and data gathered from satellites, radar, and aircraft that release weather instruments – known as dropsondes – that collect atmospheric data as they parachute down to the ocean surface.

Federal water projects serve two primary purposes: to supply water, and to manage flood risk. And the old way of doing things has, for decades, provided water managers with a reliable guide to balancing those two purposes. It was during the previous three-year record dry period in California, 2012 to 2015, that USACE and several partners, including CW3E and other federal, state, and local agencies, began to envision a way of managing water resources that reflected the new boom-and-bust reality.

At Lake Mendocino, a substantial amount of rain had fallen and collected in the months prior to the 2012-2015 interval, and much of that water was released to make room for the next storm. But the next storm was a long time coming. Flood risk was zero – and water supply was far less than it needed to be. They were out of balance.

Water is released down the spillway from Coyote Valley Dam at Lake Mendocino near Ukiah, California. USACE PHOTO BY BRANDON BEACH

According to Cary Talbot, Ph.D., PE, USACE national lead for the FIRO program, Lake Mendocino provided a near-perfect candidate for piloting this new research and development initiative: It was small, with a flood storage capacity of 22,400 acre-feet. Compared to the state’s largest reservoir, Shasta Lake (more than 4.5 million acre-feet), it would be relatively easy to manage and measure.

The lake’s water control manual was also ripe for revision. The dam, built in 1959, was old enough to have been equipped with a nuclear fallout shelter. For decades before the dam’s construction, Talbot said, the reservoir had been fed by water transferred from the adjacent watershed, the Eel River, to generate hydropower. But the project’s output had been decreasing steadily since 2006, for several reasons, and the utility that operated it, Pacific Gas & Electric, was looking to phase it out. “Its future, in terms of the amount of water coming through that transfer, is in great doubt,” said Talbot. “So the water control manual that was written back in 1959 is tailored to a water budget that is now very different from what it was originally.”

Talbot is a hydraulic engineer and program manager at the Coastal and Hydraulics Laboratory at the Engineer Research and Development Center (ERDC), the main USACE R&D center in Vicksburg, Mississippi. Another mark in Lake Mendocino’s favor, he said, was that there was already a great deal of collaboration and communication among project stakeholders. “State, federal, and local agencies were all working together,” he said, on discussions about balancing flood risk and water supply. The National Marine Fisheries Service had already issued a biological opinion on appropriate releases from Lake Mendocino. “It was fertile ground,” said Talbot, “to try something new and different.”

In late 2014, when the mostly dry Mendocino lakebed was littered with fish skeletons and ancient boating and fishing gear, Congress appropriated an extra $2 million dollars for research in the USACE budget, and Phase One of the FIRO experiment – researching whether FIRO could be proven viable and safely implemented for the Lake Mendocino project – was underway.

How FIRO Works

At Lake Mendocino – and at the handful of California reservoirs where USACE and its partners have now begun to experiment with forecast-informed deviations from existing water control manuals – the process began by forming a steering committee composed of leaders from relevant stakeholder agencies, who composed a preliminary viability assessment (PVA). At the core of a PVA are answers to a few basic questions: Is it likely that the benefits of FIRO will outweigh any associated costs? What are some scenarios we can test here?

people in group photo along lake shoreline
USACE FIRO program personnel at Lake Mendocino in 2018. USACE PHOTO BY BRANDON BEACH

The idea behind FIRO is simple. In cross-section, a reservoir is divided into two units or “pools” of water: The flood control pool sits on top of the conservation pool, which is usually a multi-purpose allotment of water for meeting all other demands on the resource. According to Cuong Ly, senior hydrologic and hydraulic water management engineer for USACE’s South Pacific Division, the line between these pools is fixed at non-FIRO dams, a hard boundary set by the water control manual.

“Above that line, the Corps controls the resource; below that line, it’s the water agency’s job,” Ly said. “FIRO doesn’t move that line. It creates a buffer that, in most cases, allows the Corps to store more water for conservation purposes. But in rare events – say, a huge storm is forecast to be coming in – the local agency can allow the Corps to release water below the conservation line.” Knowing with near certainty that a storm is either on the way, or unlikely, allows the agencies to retain or release water accordingly.

In December 2020, after USACE and its partners had demonstrated that FIRO could yield clear benefits in the management of Lake Mendocino, they released their final viability assessment (FVA). A culmination of the steering committee’s six-year effort, the FVA provides strong support for the adoption of FIRO at Lake Mendocino. At the time of this writing, the water control manual for Lake Mendocino is being updated to permanently implement FIRO.


In practice, the easy-to-explain flexibility offered by FIRO isn’t as easily achieved: Multiple stakeholders have claims to the water in the conservation pool, and every reservoir has a unique set of social, geographic, and hydrologic variables. Even as they were initiating FIRO and documenting success at Lake Mendocino, Talbot, his colleagues at the Corps of Engineers, and their partners working at other USACE-owned dam sites launched Phase Two of their FIRO R&D initiative: the expansion of FIRO to other pilot projects that will allow them to further explore FIRO at a variety of sites and develop a screening process for taking the program nationwide.

“FIRO doesn’t move that line. It creates a buffer that, in most cases, allows the Corps to store more water for conservation purposes.” – Cuong Ly, USACE

New Bullards Bar Reservoir, in the foothills of the Sierras, is the location of one of the new pilot pro- grams underway to further explore FIRO. PHOTO BY JUSTIN SMITH VIA WIKIMEDIA

These new pilot projects were selected, in part, because they presented challenging new variables: In the foothills of the Sierra Nevada Mountains, steering committees were formed at New Bullards Bar Reservoir and Lake Oroville, northeast of Sacramento. “We wanted to go to the Sierras,” Talbot said, “where now we have snow as a factor, and two dams operating together, because their outflows come together at Yuba City.”

Just outside arid Orange County, FIRO was rolled out at Lake Prado, where the Prado Dam was built in 1939 for flood control. “Lake Mendocino is rural,” said Talbot. “Prado is absolutely urban. You’re right there, upstream from Disneyland and the concrete jungle, and below the San Bernardino Valley and some steep mountains.”

One of the new pilots – Lake Sonoma, which supplies the lower Russian River and operates hatchery programs for both steelhead trout and coho salmon – was selected for the obvious reason that it may validate the benefits of FIRO on a watershed scale. “This is the first water year we’ve applied FIRO at Lake Sonoma,” said Sing. “We had a deviation in place last year, but because it was a dry water year, there wasn’t an opportunity to test it out.”

The Widening World of Water Issues

In addition to these reservoir-specific problems, several emerging concerns are poised to have a significant influence on how USACE-owned dams are operated, especially in the arid West. The Corps of Engineers and other federal partners have begun to reckon with the flood risk associated with wildfire burn scars, for example – and in the summer of 2020, much of Lake Sonoma’s drainage area was consumed by the Walbridge Fire, a lightning-caused fire that burned about 20,000 acres up to and along the lake’s entire southwestern shore.

Aerial view of the Prado Dam, a U.S. Army Corps of Engineers structure near Corona, California. A new report published by the Forecast Informed Reservoir Operations program, or FIRO, found using advanced weather and streamflow forecasts to enhance water storage capabilities at the dam could conserve enough water to supply an additional 60,000 people per year. PHOTO COURTESY OF THE ORANGE COUNTY WATER DISTRICT

According to Malasavage, the fire evoked a prompt response from county, state, and federal partners. “There were some areas of steep terrain that we were concerned were at elevated landslide risk,” he said. Such risks, of course, would necessarily be factored into water resource decisions – but because 2021-2022 was a drought year, “those consequences never came to fruition, at the cost of having no rain.”

Another emerging water resource issue is the depletion of aquifers within watersheds that feed and are supplied by USACE-operated reservoirs. In 2020, USACE’s Institute for Water Resources released a report on Managed Aquifer Recharge (MAR), which, the authors wrote, “can augment surface storage and increase resilience of USACE projects and improve our nation’s water security.” At the time, USACE and its partners were using, or considering the use of, MAR in 17 states. “The authorities for using MAR in USACE projects are modest but increasing.”

FIRO and MAR are a natural fit, said Talbot. “We actually just had a workshop, two weeks ago, with a presentation called FIRO-MAR,” he said – and MAR is already a function at Lake Prado, where the USACE Los Angeles District’s policy is not to store water behind the dam for very long. “They, in coordination with the Orange County Water District, say, ‘We’re going to release it at such a rate that every drop can be infiltrated by Orange County into the ground.’ That’s very much a managed aquifer recharge exercise, and FIRO is directly impacting that.”


At each of the reservoirs where FIRO has been implemented, it has demonstrated benefits even during this transitional Phase Two, said Ly. “We haven’t formalized an update of a water control manual yet. At Mendocino and Prado Dam, we’re implementing the deviation process, which is a temporary process, while we update the manual. But we’re already realizing the benefit of FIRO.”

These successes have led USACE and its partners, while they finalize the screening tool they’ll use to select more projects, to adopt an ambitious vision for FIRO’s future. “Applying that screening process to our portfolio of dams,” Talbot said, “is our Phase Three. That’s what we’re starting now.” One area on the radar is Oregon’s Willamette Valley – a system of dams, 13 of them owned by USACE, that operate together as a unit.

“We want to try FIRO in that kind of a setting,” said Talbot. “We’re going to learn how to do a system in the lab. We’ll also take what we learned from there and do another system in some other part of the country – not the West Coast – and see how it applies there.”

“We haven’t formalized an update of a water control manual yet. At Mendocino and Prado Dam, we’re implementing the deviation process, which is a temporary process, while we update the manual. But we’re already realizing the benefit of FIRO.” – Cuong Ly, USACE

If it applies as well as it has at a handful of California reservoirs, the nation is at the threshold of a new era in water resource management. In most of California, a prolonged drought ended during the 2022-2023 water year, during which 31 atmospheric rivers brought substantial rain throughout the state. One three-week stretch, in late December and early January, was the state’s wettest in 161 years. “There were periods of time this year when we had to release water, from both Lake Mendocino and Lake Sonoma, because the reservoir was full,” said Sing. Through the periods of dry weather that followed in early spring, FIRO was applied to retain some water. Applying FIRO at both reservoirs, Talbot said, resulted in an additional 24,000 acre-feet of stored water at the end of the season – enough for 48,000 homes.

Water flows through the Oroville Spillway at Lake Oroville in Butte County, California. CALIFORNIA DEPARTMENT OF WATER RESOURCES PHOTO

The benefits of FIRO are even more obvious when weighed against what it costs, in coastal California, to procure such large amounts of fresh water. At Lake Prado, a startling result was achieved over a single weekend late in the 2020 water year, according to Ly. “From Friday to Saturday, Orange County was able to save 30,000 acre-feet of water.”

On the coast of nearby San Diego County, a $1 billion desalination plant was built in 2015 to provide fresh water to the community. “In an entire year,” Ly said, “they produce about 30,000 acre-feet.” The annual cost of producing that much fresh water, according to the Wall Street Journal, is about $49 to $59 million.

At Lake Prado, and at other reservoirs using FIRO, the cost of saving that water, after a new water control manual is in operation, will be virtually zero.

“We’re just taking our existing infrastructure and operating it better,” said Talbot. “There’s no structural change. We’re just spending a couple million to investigate how to manage it better. And it’s proven itself in both wet years and dry years. That’s given us a lot more confidence as an agency that this is the right way to go.”

This article originally featured in the 2023-2024 edition of America’s Engineers: The People, Programs, and Projects of the U.S. Army Corps of Engineers


Subscribe to the America's Engineers newsletter and never miss out on any of the recent stories about the incredible people, programs, and projects of the U.S. Army Corps of Engineers.

You have successfully subscribed to the newsletter

There was an error while trying to send your request. Please try again.

America's Engineers will use the information you provide on this form to be in touch with you and to provide updates and marketing.