Flood Costs Beyond Buildings: Lessons From Lake Champlain

Floods account for more than 75 percent of federal disaster declarations and cost more than any other natural disaster. Early warning has reduced deaths, but costs keep rising as flood-prone land is developed. Standard flood-loss models focus on damage to buildings — but a U.S. Geological Survey case study at Lake Champlain shows that the other costs of a flood can be far larger.

Lake Champlain and its watershed in New York, Vermont and Quebec. Modified from Flynn and Hayes (2016). USGS.

The lake and its floods

• Lake Champlain covers 436 square miles between New York (west) and Vermont (east); its northern end crosses into Canada, within the 9,277-square-mile Lake Champlain Basin. Its only outlet is the Richelieu River, flowing north into Quebec.
• Over the past 50 years, building on both sides of the border has drained wetlands, added pavement and straightened tributaries, sending more runoff into the lake.
• Flooding begins at 99.57 feet above sea level; major flooding at 101.07 feet.
• In spring 2011, after the third-deepest snowfall season in over 100 years, almost no gradual snowmelt, and then heavy rain, the lake hit a record 102.77 feet — more than a foot above the 1993 record — growing by 66 square miles (about 5.8 percent). With a single outlet, it took about six weeks to peak and drop back below flood stage. Waves eroded shorelines up to 5 feet above the still-water level in places. (The record is often given as 103.07 or 103.27 feet at Burlington, because of different measurement baselines and wave action.)

The study

The International Joint Commission, which handles U.S.–Canada boundary waters, set up the Lake Champlain–Richelieu River (LCRR) Study Board to assess flood risk and weigh ways to reduce it. USGS economists estimated U.S.-side costs of lake-rise flooding — slow-moving water with wave action (not tributary floods) — as inputs to the board's cost-benefit models. In those models, a submerged weir in the Richelieu River came out with benefits exceeding costs, while dams or a new Canadian canal were never considered feasible or appropriate.

FEMA's Hazus model estimates damage to buildings. The board added Principal Indicators (PIs) for other costs. USGS economists estimated six:

Flood damage to the Cobble Ridge Road bridge after heavy rain in Londonderry, Vermont. Photograph by the U.S. Geological Survey.

What drives the numbers

Lodging. Lake floodwater moves very slowly and rarely cuts off utilities; no permanent homes are destroyed, because homes that would take 8 feet of water above the first floor were lost decades ago. But water wicks up walls at 12 inches deep and reaches outlets at 18 inches, forcing long stays away for repairs — and people expecting to be out 90 days or more rent rather than stay in hotels.

Debris. Only mobile homes with more than a foot of water inside are treated as destroyed. Few lakeside homes have basements (federal flood insurance won't cover basements in new homes) — 8.72% in New York and 1.27% in Vermont — but those produce more debris.

Businesses. Hazus flagged 113 New York and 109 Vermont sites; after removing misidentified sites and those counted elsewhere, a nearly 106-foot flood would affect 5 in New York and 11 in Vermont. Businesses on docks over the water had the highest, steadiest losses.

Recreation. Campgrounds and marinas were sorted by how much of their property a 2011-level flood would submerge. Revenue varies sharply over the five-month season — slow before Victoria Day and Memorial Day, often full in July and August — so losses depend on when waters recede.

Roads. Slow water rarely badly damages bridges or scours roads; studies abroad suggest repairs often cost only 1 to 3.5 percent of full replacement. Rail lines need spot checks only at floods at least a foot above 2011's record.

Water. Only two lakeside plants face serious problems above the 2011 level — though the Vermont one serves 40,000 customers. Slowly rising water leaves time to respond, and the lake is so large that sewage spills won't reach water intakes far offshore.

Beyond the buildings

• At a moderate flood (101.07 ft), where vulnerable buildings are long gone and new building is generally banned, the non-building costs are more than 4 times building damage — more than 18 times counting roads.
• At about the 2011 record, they are still substantial: about 40 percent of building damage without roads and just under 90 percent with them.
• At an extreme 105.57 ft, many buildings thought to be outside any flood zone would be hit, and the ratios fall to about 10 and under 20 percent.

The lesson: for moderate floods, costs beyond buildings can be most of the bill.

A two-day rainstorm in July 2023 sent boulders down a hillside in Plymouth, Vermont, burying a resident's car and moving a hot tub. Photograph by the U.S. Geological Survey.

Sources

Based on U.S. Geological Survey Fact Sheet 2023–3034, "Flood damage costs beyond buildings—A Lake Champlain case study," summarizing Rhodes (2022) and the International Lake Champlain–Richelieu River Study Board's reports; a work of the United States government in the public domain.