rTDataScrub

This repository aims to build a Hidden Markov Model that identifies errors in stream temperature data. By providing time series of observed water temperature data and summary annual air temperature statistics (e.g., mean, range) derived from climate models such as PRISM or ClimateBC, this model will probablistically parse data into either an air or water state. By providing probabilities, the model offers a more objective consideration of the data that can then be assessed in context. It also can leverage data from multiple sites and consider the contribution of those data in it's state estimates. Ultimately, this model could be developed into a tool that provides anyone collecting stream temperature data with a simple way of identifying errors.