Exceedance probability is the likelihood that a specific value, such as a flood depth or a financial loss, will be exceeded within a given time period. A 1% annual exceedance probability means there is a 1-in-100 chance of that event occurring in any single year. Despite being straightforward in concept, exceedance probability is widely misunderstood, routinely confused with return periods, and central to how flood risk, insurance pricing, and climate risk assessments work.
This article covers the math behind exceedance probability, explains how to read EP curves, and shows how the concept connects flood hazard mapping to insurance pricing and forward-looking climate risk.
What Is Exceedance Probability?
Exceedance probability answers a specific question: what is the chance that a loss or event magnitude will exceed a given threshold? It is the complement of the cumulative distribution function (CDF). Where a CDF tells you the probability of being at or below a value, exceedance probability tells you the probability of being above it.
In risk management, exceedance probability is used to express statements like “there is a 4% chance that flood damage to this building will exceed $500,000 in any given year.” The threshold ($500,000) and the probability (4%) together define a point on what is called an exceedance probability curve.
The concept applies across domains. Hydrologists use it to classify floods. Insurers use it to price catastrophe risk. Climate scientists use it to project how hazard frequency shifts under warming scenarios. In each case, the core math is the same.
Annual Exceedance Probability vs Return Period
Annual exceedance probability (AEP) and return period express the same concept in different units. A 1% AEP flood has a return period of 100 years. A 4% AEP flood has a return period of 25 years. The conversion is simple: return period = 1 / AEP.
| Annual Exceedance Probability | Return Period | Common Name |
|---|---|---|
| 50% | 2 years | Frequent flood |
| 20% | 5 years | Minor flood |
| 10% | 10 years | Moderate flood |
| 4% | 25 years | Significant flood |
| 2% | 50 years | Severe flood |
| 1% | 100 years | FEMA Special Flood Hazard Area |
| 0.2% | 500 years | FEMA Moderate Flood Hazard Area |
The USGS now recommends using annual exceedance probability instead of return period because “100-year flood” misleads people into thinking the event cannot recur for another century. In reality, a 1% AEP flood can happen in consecutive years. The probability resets each year, the same way a coin flip has a 50% chance of heads regardless of what came before.
How to Calculate Exceedance Probability
The formula for calculating the probability of at least one exceedance over multiple years is:
P = 1 – (1 – AEP)n
Where P is the probability of at least one event, AEP is the annual exceedance probability, and n is the number of years. For a 1% AEP flood over a 30-year mortgage, that calculation is P = 1 – (1 – 0.01)30 = 0.26, or a 26% chance of flooding at least once during the loan term. One in four mortgages in a 1% AEP flood zone will experience flooding.
The example above used a 1% AEP flood over a 30-year mortgage. Use the calculator to check any combination of event probability and time horizon.
| AEP | Over 10 Years | Over 20 Years | Over 30 Years | Over 50 Years |
|---|---|---|---|---|
| 10% (10-yr) | 65% | 88% | 96% | 99% |
| 4% (25-yr) | 34% | 56% | 71% | 87% |
| 1% (100-yr) | 10% | 18% | 26% | 39% |
| 0.2% (500-yr) | 2% | 4% | 6% | 10% |
The table shows why the “100-year flood” framing is dangerous. A building with a 1% AEP has a 26% chance of flooding during a 30-year mortgage and a 39% chance over 50 years. Framing it as “1% annual exceedance probability” makes the ongoing risk clearer than calling it a “100-year flood.”
What Is an Exceedance Probability Curve?
An exceedance probability curve (EP curve) plots loss amounts against the probability of exceeding each amount. The horizontal axis shows the loss in dollars. The vertical axis shows the annual probability that actual losses will exceed that dollar amount. The curve slopes downward from left to right: small losses have high exceedance probabilities (they happen often), while catastrophic losses have low exceedance probabilities (they happen rarely).
EP curves are generated from catastrophe model simulations. A typical model runs 10,000 to 100,000 synthetic years of weather events, computes the damage from each event using depth-damage functions, and ranks all the losses from smallest to largest. Each ranked loss gets an associated probability, and plotting them creates the EP curve.
The shape of the EP curve reveals the risk profile. A portfolio exposed mainly to frequent small events (like thunderstorms) produces a steep curve that drops quickly. A portfolio exposed to rare catastrophic events (like major earthquakes) produces a flat curve with a long right tail. Insurers and risk managers use the shape to understand whether risk is concentrated in tail events or spread across frequent occurrences.
OEP vs AEP: Two Types of EP Curves
There are two standard EP curve types, and they answer different questions:
| Attribute | OEP (Occurrence) | AEP (Aggregate) |
|---|---|---|
| Full name | Occurrence Exceedance Probability | Aggregate Exceedance Probability |
| What it measures | Largest single event loss in a year | Total of all event losses in a year |
| Counts per year | Only the peak event | All events summed |
| Curve position | Always below AEP curve | Always above OEP curve |
| Primary use | Reinsurance pricing, PML | Portfolio pricing, AAL, capital |
| Related metrics | VaR (Value at Risk) | TVaR (Tail Value at Risk) |
The OEP curve focuses on the single worst event in each simulated year, making it the right tool for sizing probable maximum loss (PML) and structuring per-occurrence reinsurance. The AEP curve sums all events, making it the right tool for annual budgeting, portfolio pricing, and computing climate value at risk. The area under the AEP curve equals the average annual loss (AAL), which is the expected long-run annual cost of catastrophe events.
Exceedance Probability in Flood Risk and Insurance
Exceedance probability is embedded in US flood regulation. FEMA defines flood zones using AEP thresholds: the Special Flood Hazard Area (SFHA) is the 1% AEP zone, and the Moderate Flood Hazard Area is the 0.2% AEP zone. Any federally backed mortgage on a property inside the 1% AEP zone requires mandatory flood insurance through the National Flood Insurance Program.
In catastrophe modeling, EP curves are the primary output. Insurers use them to set premiums, structure reinsurance layers, and satisfy regulatory capital requirements. The National Association of Insurance Commissioners (NAIC) requires catastrophe model output, including EP curves and average annual loss, for property insurance ratemaking in catastrophe-prone states.
The connection between exceedance probability and damage estimation is direct. At each return period (10-year, 25-year, 100-year, etc.), a flood model produces a flood depth at a given location. That depth feeds into depth-damage functions, like FEMA’s HAZUS curves (196 individual curves across 33 building types), which convert depth into a damage ratio. Multiplying the damage ratio by the building’s replacement value gives the loss at that return period. Plotting losses against their associated exceedance probabilities across all return periods produces the EP curve. For commercial flood damage, the framework uses 10 occupancy-specific curves covering retail, warehouse, office, hospital, and six other categories — each with a distinct depth-damage profile.
Exceedance Probability in Climate Risk Assessment
Traditional EP curves assume a stationary climate: the probability of a 1% AEP flood stays at 1% indefinitely. Climate change breaks this assumption. As temperatures rise, atmospheric moisture increases (roughly 7% per degree of warming via the Clausius-Clapeyron relationship), rainfall intensity grows, and what was historically a 1% AEP flood becomes more frequent. A flood with a 1% AEP under today’s climate might have a 2% or 4% AEP by 2050 under high-emissions scenarios.
Forward-looking exceedance probability is central to TCFD and ISSB climate disclosure frameworks. Companies reporting physical risk exposure need to show how EP shifts under different climate scenarios (such as SSP2-4.5 and SSP5-8.5) and time horizons (2030, 2040, 2050). This requires combining climate projection models with depth-damage curves to build scenario-specific EP curves for each asset.
Continuuiti’s damage estimation service does this by applying both HAZUS and JRC depth-damage curves across return periods for any location globally, producing dual-source loss estimates that feed into portfolio-level exceedance probability and climate value at risk calculations.
Frequently Asked Questions
What is a 1% annual exceedance probability?
A 1% annual exceedance probability (AEP) means there is a 1-in-100 chance of an event of that magnitude or greater occurring in any given year. It is equivalent to a 100-year return period. Over a 30-year mortgage, there is a 26% chance of at least one 1% AEP flood occurring. FEMA uses the 1% AEP threshold to define Special Flood Hazard Areas where flood insurance is mandatory.
How do you calculate exceedance probability?
The formula is P = 1 – (1 – AEP)n, where AEP is the annual exceedance probability and n is the number of years. For a 1% AEP event over 30 years: P = 1 – (1 – 0.01)30 = 0.26, or 26%. This gives the probability of at least one exceedance over the time period.
What is the difference between OEP and AEP?
OEP (Occurrence Exceedance Probability) measures the probability of the single largest event loss in a year exceeding a threshold. AEP (Aggregate Exceedance Probability) measures the probability of total annual losses from all events exceeding a threshold. OEP is used for reinsurance pricing and probable maximum loss. AEP is used for portfolio pricing, average annual loss, and capital allocation.
What is the probability of a 1000-year flood?
A 1000-year flood has a 0.1% annual exceedance probability, meaning a 1-in-1000 chance of occurring in any given year. Over 50 years, the probability of at least one such event is about 5%. The “1000-year” label does not mean the flood happens once every 1000 years. It can happen in consecutive years because the probability resets annually.
How does climate change affect exceedance probability?
Climate change increases exceedance probability by intensifying rainfall and raising flood depths. A flood that historically had a 1% AEP may shift to 2% or 4% AEP under warming scenarios by 2050. Infrastructure designed for historical exceedance probabilities becomes under-designed for future conditions, which is why forward-looking climate risk assessments recalculate EP curves under multiple emission scenarios.
Exceedance probability underpins every quantitative conversation about flood risk, insurance pricing, and climate exposure. Whether a bank is screening a mortgage portfolio for flood risk, an insurer is pricing a catastrophe layer, or a sustainability team is reporting physical risk under TCFD, the EP framework translates raw hazard data into actionable financial metrics. The math is straightforward. The challenge is getting the inputs right: accurate flood depths, appropriate depth-damage curves, and climate projections that reflect how probabilities shift over time.
