SEPA modelled that with a computer model. I think that members all have a copy of my submission, which shows a map of Dounie, in the Sound of Jura, where I live. The purple area shows the existing model that SEPA used to model where the waste—the solids and the emamectin benzoate—will go. The model suggested that it is all underneath the farm, but 99 per cent is swept away—as soon as it leaves the black square on the map the model does not care where it goes; it is gone, as far as the model is concerned.
SEPA then did a test with the new model, which showed that waste goes elsewhere. Not all of it goes elsewhere, but 86 per cent leaves that square and the model still does not care where it goes. The amount that goes into the red area on the map is a kilogram of waste per square metre per year, with the emamectin bound in—it is excreted by the fish, so it sits on the seabed.
I am saying that the old model that is still in use, and has been used for 15 years to do all the pollution permissions under the controlled activities regulations, is inadequate. It is a flawed model that ignores the fact that the sea bed slopes. The reason why the red and purple areas on the map are different is that the sea bed slopes there, but the old model does not account for that. The industry is expanding into places where there are fast currents, steeper slopes and more complicated bathymetry, so the current model is inadequate.
There is a new model that is better, but it still exports 86 per cent of waste from the site: that waste is then ignored, as if it had gone away. It has not gone away. It has gone somewhere else: it settles up the coast. Another fish farm 1km away will get that 86 per cent of the waste from the fish farm that we are looking at. That impact is not in the model.
I just spoke to Anne Anderson from SEPA; she does consenting and compliance. We wrote to her in October. Apparently SEPA has replied—I have not seen the letter, but apparently it has been posted—saying that it will use the new model instead of the old one. However, the new model is opaque. It has not been peer reviewed and no one knows what assumptions are built into it. The new model should be publicly and independently scrutinised, because the assumptions in computer models determine outcomes, and the model has the enabling of industry expansion as one of its goals. That is not a good basis for setting levels for anything, and—