Most of our models are specific to predicting concentrations in water bodies, and those models have been developed in concert with the U.S. EPA. Over the course of the years, these models have been refined. There's been what they call science advisory panels that have been held to validate some of those models, and they are models that are being used jointly. As we've progressed from one model to the other, we've validated that the results are consistent with what we were expecting.
The models are used as an initial lower-tiered estimate of what animals or fish or insects would be exposed to, and then if those show significant risk, if we have good monitoring data, then we will turn to that to see whether or not the predicted model outcomes are similar to what we're seeing in the environment. If they are, then that validates our assumptions in terms of risk. If they're not, that's when we need to do more work to see whether or not the models are overly conservative and whether the true real-world data is actually telling us a different story.
When we have good, solid monitoring information, as far as we're concerned, that's where the bar is set. The sad thing is that we often don't have that very good, strong, robust monitoring data.