The discussers’ criticism focuses on the validity of the literature models selected by the writers, and not on the analysis performed in the paper with the use of selected models. The discussers consider the analysis in the paper incomplete as the predictions of the literature models used are not in agreement with those of CORMIX, a USEPA-approved computer model. However, the models used in the paper are well established in literature, and their results are carefully validated with the Visual Plume (VP), a widely used USEPA computer model (Frick et al. 2003). Based on the above, the underlined issue raised by the discussers is the differences in the predictions between CORMIX on one hand, and the VP and other well-established models on the other. It appears rather odd that the discussers, who happen to be the developers of CORMIX, direct their criticism against the users of alternative models; this is hardly the most effective way to resolve the plume modeling validation issue. Nonetheless, the writers feel compelled to defend their selection of the literature models used in the paper for analysis and validation, and for this reason they note the following. The discussers state that the analysis is incomplete, elaborating further that the relations of the models selected and used in the paper, as well as in the VP UM3 and VP NRFIELD models used for validating the results, do not consider the gravitational density current flow that develops after initial mixing. They also state that gravitational spreading and mixing is absent from most mixingzone models, “with one important exception—the USEPAapproved CORMIX system.” To prove the importance of these mechanisms, the discussers refer to their previous discussion (Jirka and Doneker 2003) of field dye tests and modeling of the Ipanema Beach outfall (Carvalho et al. 2001), in which density current mixing appears to be important. It is interesting to note that in the above paper (Carvalho et al. 2001) and its closure (Roberts and Carvalho 2003), the predictions of RSB, UM3, and CORMIX were compared with field test results and it was concluded that “with suitable assumptions, all the models reasonably predict near field dilution,” and that “CORMIX grossly underestimates the observed thicknesses” of the spreading layer. Furthermore, the discussers’ statement that the models do not include gravitational spreading is incorrect for NRFIELD. NRFIELD does include spreading and the mixing in the spreading layer, as dealt with at length in Roberts and Carvalho (2003). One should recognize that plume mixing involves complex physical phenomena, not all of which can be reliably modeled at
Read full abstract