Floods account for more than 75 percent of Federal disaster declarations and lead other natural disasters in economic costs. Early-warning systems have lowered flood-related fatalities, but costs continue to rise as flood-prone areas continue to be urbanized (U.S. Geological Survey, 2006). A Lake Champlain case study shows that at moderate flood heights, the economic costs of non-structural damages or losses—such as temporary lodging, residential debris removal, commercial revenue losses, and road repair—can be greater than economic damages to buildings. For unprecedented flood heights, non-structural damages can still total more than 10 percent of structural damage costs.

Lake Champlain Case Study Context and Approach

Lake Champlain is a 436 square-mile lake bordered by New York on the western side and Vermont on the eastern side (fig. 1). The lake’s uppermost region spans the U.S.-Canadian border and sits within the 9,277 square-mile Lake Champlain Basin. Lake Champlain’s only drainage outlet is the Richelieu River, which flows north into the Canadian province of Quebec. Increasingly over the past 50 years, housing and other construction on both sides of the international border have converted wetlands, increased impervious surface area, and straightened tributaries, increasing flows of rain-water run-off. Occasionally, flooding causes significant economic damages in New York, Vermont, and Quebec, with the highest floods on record in the early 1990s and in 2011. Flooding on Lake Champlain begins at 99.57 feet (ft) above sea level, and major flooding at 101.07 ft above sea level. Herein, flood “depth,” “height,” and “elevation” all refer to water levels between 99 to 106 ft above sea level.

Following the third deepest seasonal snowfall in over 100 years, almost no seasonal snowmelt, then heavy rains, the spring of 2011 delivered record flooding to 102.77 ft. This was more than 1 foot over the 1993 record and expanded the lake’s area by 66 square miles, or about 5.8 percent (Lake Champlain Basin Program, 2013; International Lake Champlain-Richelieu River Study Board, 2019). (The record flood height, 102.77 ft, is often reported as 103.07 ft or 103.27 ft in Burlington, Vermont, owing to different origin measurements for digital elevation models, and some wave action, and differs in the Lake Champlain Basin Program [2013] and International Lake Champlain-Richelieu River Study Board [2019] reports.) With a single drainage point for the extra water volume and multiple rain events, it took approximately 6 weeks for the flood to peak and then return to a lake level below flood stage. Wind-to-wave-driven erosion occurred up to 5 ft above static lake elevation in some areas.

The International Joint Commission (IJC) between the United States and Canada handles boundary water issues between the two countries. The IJC Lake Champlain Richelieu River (LCRR) Study Board project is a bi-national, multiagency effort to assess flood risk, and to project costs and benefits (including savings) from specific structural or societal interventions that may mitigate Lake Champlain flooding or flood effects.

The U.S.-side flood-based economic damages and losses summarized in this fact sheet served as inputs to new modeling tools developed for the LCRR Study Board project that calculate benefit-to-cost ratios associated with structural interventions. For example, adding a submerged weir in the Richelieu River yielded a greater-than-one benefit-to-cost ratio in late-stage modeling (International Lake Champlain-Richelieu River Study Board, 2022), whereas a dam on either side, or an entirely new canal on the Canadian side, were never entertained as cost feasible or even appropriate (International Lake Champlain-Richelieu River Study Board, 2021).

The scope of the U.S.-side economic analysis done by U.S. Geological Survey (USGS) economists pertained only to so-called lake-rise flooding that has little velocity but can have wave action. Flood damage or losses on tributaries to Lake Champlain were not included, because tributary rivers are managed by the United States only (in a bi-national project assessment). Uncommonly low lake levels compared to recent decades (95 ft and lower) were also considered as a stakeholder concern (adding a weir helps to raise minimum depths too).

Economic damages from flooding are often modeled using the Federal Emergency Management Agency’s (FEMA) Hazus-Multi-Hazard tool (https://www.fema.gov/flood-maps/products-tools/hazus). Hazus models economic costs of structural damages to buildings within the projected flood level of the census tract each building type is located in. The LCRR Study Board project used Hazus modeling to estimate structural cost damages at different flood depths, then defined Principal Indicators (PIs) to supplement, or alternatively estimate, results from applying Hazus—where gaps exist or where unmodified Hazus output values may be inaccurate in the LCRR context. The PI approach provides depth-damage estimates for a set of costs separate from estimates of building damages from lake-rise flooding. Seven PIs were estimated on the Canadian and U.S. sides of Lake Champlain, in addition to Hazus depth-damage estimates, after emergency response costs and ecosystem service losses were dropped as PIs for time and budget reasons. Depth-damage agricultural yield losses were estimated for the Canadian and U.S. sides by the Canadian economic team.

The U.S.-side economists produced cost estimates for six PIs: temporary lodging costs; residential debris clean-up and disposal; damage to roads and bridges (and railroads); damage to water treatment facilities; income loss from industrial or commercial properties; and separately and specifically recreation sector income loss. Residential damage estimates included the costs of securing emergency and longer-term lodging when residents are displaced from their homes by lake-rise flooding, and the costs of clean up and disposal of debris from residential property damage. In the public sector, costs of clean up and repair of damages to roads and bridges from lake-rise flooding were calculated, as were damages and potential revenue losses from flood mitigation measures and service reductions, where public or private water utilities are inundated by lake-rise flooding. In the commercial sector, revenue losses from business closures due to lake-rise flooding were calculated separately for businesses outside of the recreation sector, and for recreation-sector businesses: lakeside campgrounds; marinas; and ferry services (where ferries are also used for local commercial traffic). Table 1 provides a summary of low and high cost ranges for each PI, and one note for each on how the cost estimates were generated or how to interpret them.

In Rhodes (2022) depth-damage estimates are reported in increments of 1 foot or more, and indicate magnitudes of costs that comply with reasonable scenario assumptions for a small but consistent set of flood depths from 99.57 to 105.57 ft, where the highest water level is almost 3 ft above the historic maximum flood height of 102.77 ft. Developing different estimation techniques for each PI using expert consultation offers plausible, logical, reliable, and reproducible magnitudes for depth-damage cost estimates and provides a framework for each PI that can be easily modified if higher-quality information becomes available.

Following for each of the 6 PIs is a short list of features that affect modeling or results, where each pertains to both New York and Vermont, unless otherwise noted:

Estimated Principal Indicator Costs Compared to the Structural Damage Estimates for Buildings, at a Range of Flood Elevations

Table 2 presents and compares totals for non-structural and structural damage costs for the six PIs at three different flood heights for Lake Champlain: 101.07 ft is the moderate flood level (https://vem.vermont.gov/sites/demhs/files/documents/2018SHMP-HazardAssessmentInundationFloodingFluvialErosion.pdf); 103.07 ft is approximately the historic record flood height (spring 2011); and 105.57 ft is the highest flood level as projected by hydrologists on the LCRR Study Board project. The first row compares costs and revenue losses for five of the six PIs, with the largest-cost single PI, “Roads, bridges, and railroads,” as a separate total in the second row. These are followed by the full 6-PI total (non-structural damages) in the third row. Estimated building damage total costs (structural damages) appear for the same three flood elevations as for the PIs.

Below the first four rows are ratios of the 5-PI total costs to the building damage total costs at each of the flood elevations, followed by the same ratios for the 6-PI total (table 2). At a flood height of 101.07 ft, where vulnerable buildings have long since been cleared away and new construction is generally prohibited, the 5-PI ratio is more than 4 times (400 percent) the estimated building damages across New York and Vermont. Adding the roads, bridges, and railroads total to the numerator of the ratio raises the ratio to more than 18-to-1 (1,800 percent).

At 103.07 ft, the PI totals are substantial, at around 40 percent for the 5-PI ratio, and just below 90 percent of structural damage estimates for the 6-PI ratio. Whereas for a 105.57-ft flood, for example, many structures would be damaged that were believed to be outside of any flood zone. At this very improbable flood elevation, the 5-PI ratio drops to around 10 percent, and the 6-PI ratio drops to less than 20 percent.