Advances in Weather and Climate Forecast Verification

Dr Eric Gilleland is a project scientist at the National Center for Atmospheric Research (NCAR) in the Research Applications Laboratory (RAL), where he has performed research on statistical methods for forecast verification, especially spatial methods for high-resolution models, as well as applications of extreme value analysis, since 2002. Gilleland has a Ph.D. in statistics from the Colorado State University. Barbara G. Brown is a project scientist at the National Center for Atmospheric Research (NCAR) in the Research Applications Laboratory (RAL), where she has led efforts on development and application of methods for weather and climate forecast evaluation. She also has extensive experience providing tutorials on verification methods. Brown has MS degrees in environmental science (University of Virginia, 1979) and statistics (Oregon State University, 1983).

1. Introduction 2. Statistical Basis for Verification 3. Traditional verification for deterministic forecasts 4. Traditional verification for probabilistic forecasts 5. Diagnostic tools for univariate measures 6. Background, motivation, and overview of spatial forecast verification methods 7. Neighborhood methods 8. Scale separation methods 9. Location measures 10. Field deformation 11. Feature-based: identifying features, merging, matching, properties, and property comparisons 12. Feature-based: direct comparisons between matched features: CRA, MODE 13. Feature-based: spatial contingency table 14. Spatial Prediction Comparison Test, including image warping loss 15. Additional topics: hurricane forecasts, S2S, observation uncertainty