What is the formula for calculating recurrence interval?

What is the formula for calculating recurrence interval?

• A. years of record times rank
• B. years of record minus rank
• C. years of record divided by rank
• D. years of record plus rank

Answer: (C)

Recurrence interval measures how often, on average, you’d expect to see an event of a given size based on historical data. You list events from largest to smallest, giving the largest event a rank of 1, the next largest a rank of 2, and so on. If you have N years of data, a practical way to estimate how often an event of a certain size occurs is to divide the length of the record by its rank. So, the average time between events of that size is roughly N divided by the rank. This captures the idea that the biggest, rarest events (rank 1) are expected to occur less frequently, while smaller, more common events (larger rank numbers) happen more often. Multiplying, adding, or subtracting the record length by the rank doesn’t reflect this inverse relationship between event size (as captured by rank) and frequency. Some refinements use N+1 instead of N for a slightly less biased estimate, but the basic concept shown here is dividing the years of record by the event’s rank.

Recurrence interval measures how often, on average, you’d expect to see an event of a given size based on historical data. You list events from largest to smallest, giving the largest event a rank of 1, the next largest a rank of 2, and so on. If you have N years of data, a practical way to estimate how often an event of a certain size occurs is to divide the length of the record by its rank. So, the average time between events of that size is roughly N divided by the rank. This captures the idea that the biggest, rarest events (rank 1) are expected to occur less frequently, while smaller, more common events (larger rank numbers) happen more often.

Multiplying, adding, or subtracting the record length by the rank doesn’t reflect this inverse relationship between event size (as captured by rank) and frequency. Some refinements use N+1 instead of N for a slightly less biased estimate, but the basic concept shown here is dividing the years of record by the event’s rank.