EVS: Probabilistic Approach
Case study two, from the MARFC, looks at the question: "How well do the ensemble streamflow forecasts verify based on whether or not ensemble QPF are used as input?" This study looks at how streamflow forecast performance is influenced by 1) lead time, 2) QPF, 3) climatology, and 4) location at either a headwater or a downstream point. It looks at characteristics of data distribution and spread, and verification measures such as error, bias, correlation, forecast discrimination, forecast reliability, skill, and confidence.
The EVS tool is used for the the verification of probabilistic flow forecasts at points. Two point locations in the Juniata River basin are studied. Williamsburg is an upstream headwater location. Newport is a downstream location. This case study will be detailed in section 3. Section 4 will summarize.