Uncertainty is a word that you see used quite often in the assessment documents. As I mentioned in my presentation, we have a new assessment model, and one of the key things we have to address in a far better way than anything we've had in the past is the issue of uncertainty.
I can give you an example. In the past, we had to assume things. We didn't have enough information back before the moratorium on cod to say what the level of natural mortality was. In the models at that time, you had to put in an assumed value, so we put in a fixed value of 0.2. There was no measure of uncertainty around that at all. If we got that wrong, that could have a big impact on the assessment. Now we have a way to include uncertainty around estimates of natural mortality.
We also had to assume that the catch information that we provided was without error, that the number of tonnes of cod that we were told were removed was exact, and that it wasn't an underestimate. Those were the numbers that we had to use in assessment models back in the 1980s.
This model we have now doesn't require an exact measure of catch as one of its inputs. It requires what we call bounds in the catch. We need a lower and an upper bound, within which we will capture what the true catch was, including things we don't know much about like discarding, unreported fishing, and any of these issues that could be going on. We put in bounds, and the model will figure out where the catch—the most likely total catch—was within those bounds. It just gives it a constraint.
This issue of uncertainty is addressed much more rigorously in this new assessment model we have, which was custom written specifically for northern cod.