The hard-wired tolerance for error checks within dispatch methods has been changed from an absolute value of 1x10-8 to a value based on the Convergence slot of the Object. The checks are used to verify the validity of computed Spills against (Turbine) Release and Outflow. The new tolerance is expressed in the form:

The new tolerance computation is consistent with the epsilon calculation used for the convergence of iterative loops, and ensures that all values computed within a dispatch method are subject to the same convergence. This change does not affect model results, but may relax error criteria in some cases.