SCIENCE OF SEASONAL CLIMATE PREDICTION

The Shine Dome, 2-3 August 2006

International standards of long-range forecast assessment
by Andrew B Watkins, National Climate Centre, Bureau of Meteorology

Dr Andrew Watkins is a senior climatologist in the Bureau of Meteorology’s National Climate Centre. As head of the Special Projects sub group, and a member of the WMO Expert Team on Long Range
Forecast Verification, he manages the WMO Lead Centre for the Standardised Verification System for Long-Range Forecasts

When evaluating seasonal to interannual forecasts there are many and varied techniques that may be used, dependent upon what the user sees as most important for their line of work, or what is simply easiest for them to perform. While this is beneficial for the individual researcher or organisation, it can result in considerable difficulty when trying to determine the best techniques or methodologies used by models from different centres. Furthermore, inconsistencies in the way models are assessed can make it difficult to determine model usefulness to regional long range forecasters. For these reasons, the World Meteorological Organization (WMO) has defined a set of standards for the assessment of seasonal and longer range forecasts: the Standardised Verification System (SVS) for Long-Range Forecasts (LRF). The SVS-LRF recommends suitable diagnostics (eg. Relative Operating Characteristics and reliability diagrams for probabilistic forecasts; Mean Square Skill Scores for deterministic forecasts), key variables (eg. temperature at 2 metres, rainfall), key regions (eg. the Tropics), and recommended verification datasets against which assessments should be performed. To further enhance consistency, the Lead Centre for the SVS-LRF (www.bom.gov.au/wmo/lrfvs/) provides basic computing subroutines, gridded verification data and a system for displaying results. Assessment of seasonal and longer range models, following the guidelines of the SVS-LRF, is required for those National Meteorological and Hydrological Services wishing to be accredited by the WMO as a Global Producing Centre of Long Range Forecasts.