This third module addresses the performance of climate models output and the quantification of the associated uncertainties, including the creation of value-added products, such as graphics, maps and reports to communicate climate forecasts reliability and products scores.
Through the Module 3, participants will acquire theoretical and practical knowledge on current approaches on verification strategies of seasonal forecast products in the Mediterranean Region, with particular emphasis on:
- Verification of tercile forecasts, and introduction to terciles and probabilities
- Overlap with WMO SVSLRF and mention of CCl Guidance
- Attributes of good probabilistic forecasts. Scoring probabilistic forecasts. Reliability and resolution
- Software tools for verification
- Operational verification of models and consensus forecasts
The module proposes some practical exercises on Probabilistic seasonal forecast verification, including Tool Tendency diagram, linear probability score and hit score. Data analysis and manipulation are performed with open source R software.
The exercises here proposed require some R-Statistical Software Knowledge. Please consider to go through “Module R – R Statistical Software” before starting the exercises.
About Forecast Verification
Overview & Main Issues
Seasonal Forecast Models
- 3.S1 Standardized Verification System for Long-Range Forecasts
- 3.S2 Diagnostics for physical evaluation of seasonal models
- 3.S3 Metrics and diagnostics for seasonal forecasts evaluation
- 3.S4 Commission for Climatology Guidance on Verification of Seasonal Forecasts
- 3.S5 What are General Circulation Models: Q&A on Climate Models