CONTINUOUS WATER-QUALITY SURROGATE MODELS

Surrogate modeling approaches make it possible to extend the benefits of continuous water-quality data to many constituents of interest that are not easily measured directly using in-situ technologies. Surrogate regression models for constituents such as suspended sediment, nutrients, dissolved solids, and indicator bacteria establish statistical relations between in-situ measured data and discretely sampled data to provide high-frequency water-quality data in real-time with defined uncertainty. As use of in-situ instruments for data collection has become more common, so has the application of surrogate models. Examples of surrogate applications from across the Nation, many of which can be seen at https://nrtwq.usgs.gov/, demonstrate a variety of applications and uses. Efforts have been underway to establish standard methods for surrogate approaches including continuous and discrete data collection, model development and statistical considerations, model verification and update, proper documentation, model review and approval, and dissemination of information to the public. Standard methods help facilitate a consistent and streamlined national surrogate approach that meets USGS data quality requirements and allows for timely release of data computed from models.