Home >> Forums >> Stabilization of ecosystem model

Stabilization of ecosystem model

Dear Sir/Ma'am,

I have been recently introduced to OSMOSE ecosystem model. In one of the technical paper (Grüss, A., et al. "Estimating age-and size-specific natural mortality rates for Gulf of Mexico red grouper (Epinephelus morio) using the ecosystem model OSMOSE-WFS." SEDAR42-AW-05. SEDAR North Charleston, SC, 2015.) which I came across, I found one concept that I had difficulty in understanding. It is with respect to the stability of ecosystem model. My query regarding the same is mentioned below:

Query 1: While allowing the OSMOSE model to run after calibration, how do we know that the model has attained stability?

I apologize in advance as this query might appear trite or even  pretty basic. I would like to thank you for providing such a platform to clarify doubts related to OSMOSE.

Kind regards,

Tarun Joseph


Hello Joseph,

Calibration is a very important step in the application of the Osmose model to a given ecosystem. Some key parameters defined in the Osmose model (such as larval mortality, plankton accessibility) are not well known even totally unknown and they need to be estimated by the calibration process. You can have more details about the calibration process of Osmose (i.e. the evolutionary algorithm used) by reading the publications "calibrar: an R package for fitting complex ecological models" and "A sequential approach to calibrate ecosystem models with multiple time series data" of Ricardo Oliveros Ramos.

Concerning your second question, when the calibration step is done , you will run Osmose and face the model outputs with observed data. You can consider that your model is stable when all the species integrated reach an equilibrium and simulated values (biomass, catches etc) are in an acceptable range (i.e. minimum and maximum possible values)

Best regards,

 Thank you Fabien for the reply.  In your response you have mentioned that the model is considered to be stable when all the species integrated into the model reach an equilibrium. Is it the same as telling that the model can be considered stable only when the fluctuations in simulated values (like biomass, catches, etc) across multiple time-steps become minimum given that their values lie within the acceptable range? Kindly do correct me if I'm wrong.

Best regards,

Tarun Joseph


Yes, we usually consider that the model is numerically stable when for a few decades the annual biomass/catches fluctate around values that you defined as acceptable ranges.

Basically you do no want any species to collapse or to blow up. Whether or not the stabilized values is within the expected target biomass or catches is the next step : you will need to calibrate the model in order to get a satisfactory result. It means you will need to find an a set of input parameters that give you a satisfactory solution.

Cheers, Philippe

Thank you Philippe for clarifying my query