home  /  insights  /  overtopping-spillway-capacity-and-the-design-flood
Geotechnical & Foundations

Overtopping, spillway capacity and the design flood question

An overtopping failure is rarely just a big storm. It is a comparison between the flood that arrived, the outlet capacity that was available, and the design basis the structure was built to.

July 30, 2026 · 7 min read

The short answer

An overtopping failure is rarely just a big storm; it is a comparison between the flood that arrived, the outlet capacity that was available to pass it, and the design basis the dam or levee was built to. Overtopping is the failure mechanism most often described as an act of nature, and the one where that description most often fails to survive analysis. Water going over a crest is a hydrologic event meeting a structural asset, so the question is not how large the storm was in isolation but how it compared with the flood the structure was designed and regulated to pass. That comparison brings in the spillway as built, the freeboard as it existed on the day, and the hazard classification that set the design standard in the first place.

What this article establishes

  • Whether water reaches the crest of an embankment dam depends on three quantities: the inflow hydrograph, the discharge capacity available to pass it, and the storage between normal pool and crest. Any of the three can be the deficient one, and each is established by different evidence.
  • Dam safety practice scales the inflow design flood to hazard potential, classified by the consequences of failure rather than by the size of the dam, and downstream development can convert a low-hazard dam into a high-hazard one without a shovel touching the embankment.
  • Rated spillway capacity is a design number, while delivered capacity on the day is a field condition; because debris blockage is foreseeable, an overtopping analysis should use the effective spillway capacity rather than the nominal rating.
  • Freeboard is consumed by embankment settlement, reservoir sedimentation, crest maintenance, rutting and unauthorized excavation, and finally wave run-up during the event; a survey of the actual crest profile is often the single most useful measurement because overtopping begins at the low point, not at the average.
  • Overtopping erosion works from the outside in, while internal erosion leaves a defined channel, a discrete exit point and turbid discharge before the breach; the two can coincide, and ruling out an internally weakened embankment takes evidence, not assumption.
  • An overtopping opinion that states each input, its source and how much the conclusion moves if that input changes will hold up; a single hydrologic number without its assumptions will not.

Why is dam overtopping treated as a comparison rather than a single event?

Dam overtopping is treated as a comparison because whether water reaches the crest of an embankment dam depends on three quantities, not on the storm alone: the inflow hydrograph, the discharge capacity available to pass it, and the storage between normal pool and crest. Any of the three can be the deficient one, and each is established by different evidence.

An embankment dam has no armored crest. Once flow passes over the crest of an embankment dam, erosion of the downstream face begins almost immediately, and the embankment dam is in a race it usually loses. Forensic work on an overtopping failure therefore concentrates on why water reached the crest at all.

How does hazard classification determine the design flood a dam must pass?

Dam safety practice does not set one design flood for all dams: it scales the inflow design flood to hazard potential, which is classified by the consequences of failure rather than by the size of the dam. Low, significant and high hazard classifications carry progressively more demanding requirements, with the probable maximum flood at the top of the range for high-hazard dams.

Dam hazard classification is not fixed at construction. Downstream development converts a low-hazard dam into a high-hazard one without a shovel touching the embankment. Whether the dam’s hazard classification was reviewed, and whether the spillway was reassessed against the standard that the hazard classification implied, is frequently the substantive issue in an overtopping failure investigation.

Why can a dam spillway pass less flow than its rated capacity?

A dam spillway can pass less flow than its rated capacity because rated capacity is a design number, while the capacity a spillway delivers on the day is a field condition. Sediment accumulation in an approach channel, woody vegetation on and around the spillway, an undersized or partially collapsed conduit, and debris racking across an inlet all reduce spillway discharge below the rating.

Debris blockage of a spillway deserves particular attention because it is foreseeable and treated as such in dam safety guidance. A spillway that cannot pass rated flow with the debris load its watershed reliably delivers has an effective capacity below its nominal one, and an overtopping analysis should use the effective spillway capacity rather than the nominal figure.

How are gate operations judged when a gated dam overtops?

Gate operations at a gated dam that overtopped are normally judged against the written operating or water control plan, not against hindsight. Where releases depend on gates, the hydrologic question of whether the dam could pass the flood acquires an operational one: whether the available release capacity was actually used. Gate position logs, SCADA records, power and backup power availability, access to the dam during the event, and the operating plan in force all bear on that question.

A gate release sequence that followed the written operating plan and still resulted in overtopping points at the operating plan or at the dam itself. A departure from the written operating plan points elsewhere.

How does a dam lose its freeboard?

A dam loses freeboard, the margin between maximum water surface and crest, through several slow processes over time and through wave run-up during the flood itself. Long-term settlement of the embankment and its foundation lowers the crest. Reservoir sedimentation raises the starting pool. Crest maintenance, rutting and unauthorized excavation take localized bites out of the crest.

Wave run-up, driven by wind fetch and duration, consumes the rest of a dam’s freeboard during the flood event itself. A crest survey establishing the actual crest profile, rather than the design elevation, is often the single most useful measurement available in an overtopping investigation, because overtopping begins at the low point of the crest and not at the average.

How is the flood hydrology reconstructed after a dam overtops?

The flood hydrology after a dam overtops is reconstructed as an inflow hydrograph, supported by precipitation from gauge networks and calibrated radar, stream gauge records, upstream reservoir operations and stage recordings. Where instrumentation is thin, high-water marks surveyed promptly constrain peak stage independently.

The output that matters from a reconstructed flood hydrology is the exceedance probability of what actually occurred, expressed with its uncertainty, set beside the design event. That comparison answers the argument that the flood exceeded the design basis, and the comparison should be made explicitly rather than asserted.

How is overtopping erosion told apart from internal erosion in a breached embankment?

Overtopping erosion and internal erosion leave different signatures in a breached embankment: overtopping erosion works from the outside in, while internal erosion leaves a defined channel or sinkhole-like depression, a discrete exit point, and turbid discharge before the breach. Overtopping erosion typically begins at the downstream crest edge or a low point, develops a headcut that migrates upstream through the crest, and widens as the breach deepens. Overtopping damage on the downstream face and abutment groins is broad rather than channelized.

Overtopping erosion and internal erosion can coincide, and an embankment weakened internally may fail at a flow that would otherwise have been survivable. Ruling out that sequence, in which internal weakening lets an embankment fail at an otherwise survivable flow, takes evidence, not assumption.

How is levee overtopping judged differently from dam overtopping?

Levee overtopping is judged against a different standard because levees are designed to a level of protection rather than to a maximum credible event, and overtopping is an acknowledged possibility in a levee’s design basis. That acknowledged possibility does not end the inquiry into a levee overtopping. Crown elevation relative to design grade, the condition of any crown or landside armoring, transitions at closure structures and penetrations, and prior accreditation or certification records all bear on whether the levee performed as its design contemplated at the stage it saw.

Where are expert opinions on overtopping failures challenged?

Expert opinions on overtopping failures are commonly challenged on the grounds that the reconstructed inflow relied on regional rainfall rather than measurements over the actual watershed; that spillway capacity was taken from the rating rather than the as-found condition; that freeboard used design crest elevation rather than a survey; and that hazard classification was applied as of failure rather than as of the last required review.

Overtopping analysis that states each input, its source and how much the conclusion moves if that input changes will hold up. A single hydrologic number without its assumptions will not.

This article is general technical orientation, not a failure analysis, an engineering opinion, or advice on any specific matter. Determining the cause of a particular incident requires hands-on examination by a credentialed expert.

For informational purposes only. Not engineering or legal advice, and not an opinion on the cause of any specific failure or on the conduct of any party.

Related

The practice area

failure-analysis assistanttriage · not a substitute for an expert
Happy to. Tell me what failed, how it failed, and whether the failed part and the scene are still preserved. That last one often decides what can still be established.