I am trying to fit data I recently collected and occasionally I am getting negative sigma squared values. Could anyone tell me what this is indicative of: am I using the wrong model or are my initial guesses of some parameters way off?

Robert