scope:

PURPOSE OF THIS MANUAL OF PRACTICE

The primary objective of this Manual of Practice is to guide watershed and receiving water model selection, synoptic data collection and compilation, calibration, and independent testing to define uncertainty for the users of the largely ad hoc models for determining TMDLs to manage point and nonpoint source pollution in the United States. Unlike the waste load allocation process for control of PS pollution, the United States has not holistically planned and developed all the specialized models or the technical guidance necessary for professionals to reliably develop TMDLs for all impaired waters. As a result, the development of TMDL modeling has been somewhat ad hoc and erratic, and in the last 15 years, much of the ad hoc developments were decentralized to the USEPA regions and state agencies. The exceptional development and support of a few models do not cover the full needs of TMDL modeling or correct the incomplete development of a comprehensive modeling and analysis framework. The sporadic development of simplified, unsupported TMDL models in response to unrealistic federal court decrees is especially problematic. Furthermore, many TMDL determinations use models developed by other agencies for different purposes, such as soil and water conservation or floodplain analysis. Thus, the models available to TMDL analysts are a patchwork covering some modeling requirements but leaving some determinations to rely on models developed for very different purposes, particularly those models developed for research or not developed into well-documented, practical models.

This manual of practice is necessary because too many TMDL determinations fail to select and use the most appropriate models of cause and effect for PS and NPS control of pollutants and water quality impairments. Monitoring and statistical analyses and other simple methods are selected too frequently despite the fact that many of these empirical, heuristic methods have not been established to predict reliably that load reduction strategies will correct and prevent future water quality impairments. The lack of clear guidance on parsimonious, reliable model selection and the lack of well-defined data requirements for determining cause and effect seem to be the reasons for some states and other jurisdictions resorting to unsuitable models and the expense of trial-anderror approaches under the banner of adaptive management. More important, hypothesis testing of calibration and confirmation (validation) has not been applied to establish that simulations or estimates of TMDLs are reliable.