Integrating Water Quality Standard Setting with Adaptive Implementation of TMDL Plans

The National Academy of Sciences report, “Assessing the TMDL Approach to Water Quality Management” argued that available water quality models and data to implement these models were adequate to support an ambient water quality approach for the TMDL program (National Research Council, 2001). However, a condition of this endorsement was that the modelers should report model prediction uncertainties (error bands) and that the EPA and stakeholders should recognize prediction uncertainty when making and implementing TMDL plans. Toward this end, the committee called for a commitment to “adaptive implementation” of water quality management plans. AI was subsequently endorsed by the newly proposed EPA rule governing the TMDL program (Inside EPA, July 30, 2002). However, the relationship between model prediction uncertainty and the setting water quality standards was not addressed by the NRC committee. In this paper, I will 1) explain the logic for adaptive implementation, and 2) discuss the implications of adaptive implementation for the setting of water quality standards. In both cases I will make the case for engaging stakeholders in the model development and utilization process.