Advances in Weather and Climate Forecast Verification is a resource for researchers and practitioners who are applying traditional and new methods of forecast evaluation. It offers information that will make it possible for researchers and others to apply the newest forecast methods; rather than searching for information in each individual method, providing the information needed to apply all of the techniques in one place. Selection of the appropriate method to apply to a particular problem is facilitated by comparisons of the methods and their results, which will be included in the book.…mehr
Advances in Weather and Climate Forecast Verification is a resource for researchers and practitioners who are applying traditional and new methods of forecast evaluation. It offers information that will make it possible for researchers and others to apply the newest forecast methods; rather than searching for information in each individual method, providing the information needed to apply all of the techniques in one place. Selection of the appropriate method to apply to a particular problem is facilitated by comparisons of the methods and their results, which will be included in the book. With these skills in hand, readers can make improvements in both their modelling and the evaluation of the results.
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.
Inhaltsangabe
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