Ticket #1348 (new not assessed)

Opened 3 years ago

Possible bug in Omnivory Index

Reported by: simonel Owned by:
Priority: normal Milestone: NOT SET
Component: Ecopath Version: 6.4.1
Severity: minor Keywords:


I checked with Generic37, using Toothed Whale as target. The OI in the original model for this group is 0,325. When you add import (just adding 0.8 import and then ask to sum to 1 its diet) OI increases to 1.362.

It seems to me that the problem is related to the fact that elements of DCij (diet matrix) but not the import (for which no TL estimate can be provided) are used to calculate the average TL of preys from which calculate the variance (OI).

A consequence of this calculation is the implicit assumption that TL of import is equal to 0 . The fact a fraction of the diet is at TL so low (zero, indeed) increase a lot the variance of TL of preys, explaining the large OI for predators with import.

A simple way to correct this is to assume that import have the TL of the average TL of the other preys.

Is this, although small, a possible problem/bug?

Note: See TracTickets for help on using tickets.