Connecting science to policy using spatial-statistical models: Water quality models in practice

Spatially explicit random forest based models for water quality attributes often produce results at reach scale. However, there is a spatial mismatch between the information produced by the models at reach scale and the information required for environmental management at region scale. We have provided a framework to apply reach scale water quality parameters at region scale using mixed approach informed by policy, model peformance, and literature. We recommend using percentile values between the range of 50% and 90% to report water quality attributes for environmental management.