How to Calculate Average Recurrence Interval

The average recurrence interval (ARI) is a statistical measure used to estimate how often an event of a certain magnitude is expected to occur. It's commonly used in fields like hydrology, seismology, and risk assessment to predict the timing of future events based on historical data.

What is Average Recurrence Interval?

The average recurrence interval (ARI) represents the average time between occurrences of an event of a specified magnitude. For example, in hydrology, the ARI might represent the average time between floods of a certain size. In seismology, it might represent the average time between earthquakes of a certain magnitude.

ARI is calculated based on historical data and is used to estimate the probability of an event occurring within a given time period. It's particularly useful for risk assessment and planning purposes.

How to Calculate ARI

Calculating the average recurrence interval involves several steps:

Collect historical data on the event of interest
Sort the data by magnitude or intensity
Determine the rank of the event in question
Apply the ARI formula to calculate the interval

The most common method for calculating ARI is the Gumbel distribution method, which provides a statistical estimate of the recurrence interval based on historical data.

ARI Formula

The formula for calculating average recurrence interval (ARI) is:

ARI = (n + 1) / m

Where:

n = number of years of record
m = rank of the event in question

This formula assumes the data follows a Gumbel distribution, which is commonly used for hydrological and seismological data.

The rank (m) of an event is determined by sorting all events in descending order of magnitude and assigning each event a rank based on its position in the sorted list.

Worked Example

Let's calculate the ARI for a flood event that has a rank of 5 in a dataset with 20 years of record.

Given:

Number of years of record (n) = 20
Rank of the event (m) = 5

Calculation:

ARI = (n + 1) / m = (20 + 1) / 5 = 21 / 5 = 4.2 years

Interpretation: This means we can expect a flood of this magnitude to occur approximately every 4.2 years on average.

This example shows how ARI can be used to estimate the frequency of events based on historical data.

Interpreting ARI Results

When interpreting ARI results, it's important to consider several factors:

The quality and completeness of the historical data
The assumptions about the distribution of the data
The potential for changes in climate or other factors that could affect event frequency

ARI should be used as an estimate rather than an exact prediction, as it's based on historical patterns that may not necessarily repeat in the future.

Example ARI Interpretation
ARI (years)	Interpretation
1	Event occurs approximately once per year
2	Event occurs approximately every 2 years
5	Event occurs approximately once every 5 years
10	Event occurs approximately once every 10 years
25	Event occurs approximately once every 25 years
50	Event occurs approximately once every 50 years
100	Event occurs approximately once every 100 years

Frequently Asked Questions

What is the difference between ARI and return period?

The terms "average recurrence interval" and "return period" are often used interchangeably. Both refer to the average time between occurrences of an event of a specified magnitude. The key difference is that the return period is typically expressed in years, while the ARI can be expressed in any time unit.

How accurate is the ARI calculation?

The accuracy of ARI calculations depends on several factors, including the quality and completeness of the historical data, the assumptions about the distribution of the data, and the potential for changes in climate or other factors that could affect event frequency. ARI should be used as an estimate rather than an exact prediction.

Can ARI be used for events other than floods and earthquakes?

Yes, ARI can be used to estimate the frequency of any type of event for which historical data is available. It's commonly used in fields like hydrology, seismology, insurance, and risk assessment to predict the timing of future events based on historical patterns.