Response to Comment on 'Two statistics for evaluating parameter identifiability and error reduction'

摘要

We appreciate the interest in the identifiability and relative parameter error variance reduction statistics that we describe in Doherty and Hunt (2009). Dr. Hill’s comments, referred to here as Hill (2010), include concerns specific to our paper, but also examples and issues that extend beyond our paper to the work of others. As such, they extend to broader methodological issues which underpin our general approach to model calibration and uncertainty analysis. In this regard they are similar to those raised in a recent paper (Hill, 2008a) and in a number of scientific presentations (Hill, 2006, 2007, 2008b, 2009). We consider a discussion of these issues vital to achieving changes in approaches to parameter estimation and uncertainty analysis that we believe will substantially enhance the ability of the modeling community to address societally relevant problems. We hope that discussion of the topics raised in Hill (2010) will increase the understanding of the approach that we are trying to advance, and are therefore pleased to engage in such a discussion.