作者: A. Weisheimer , T. N. Palmer
关键词: Reliability (statistics) 、 Forecast error 、 Meteorology 、 Range (statistics) 、 Precipitation 、 Scale (map) 、 Variable (computer science) 、 Environmental science 、 Consensus forecast 、 Probabilistic logic
摘要: Seasonal climate forecasts are being used increasingly across a range of application sectors. A recent UK governmental report asked: how good seasonal on scale 1–5 (where 5 is very good), and can we expect them to be in 30 years time? made from ensembles integrations numerical models climate. We argue that ‘goodness’ should assessed first foremost terms the probabilistic reliability these ensemble-based forecasts; reliable inputs essential for any forecast-based decision-making. propose ‘5’ reserved systems not only overall, but where, particular, small ensemble spread indicator low forecast error. study regional temperature precipitation current operational system European Centre Medium-Range Weather Forecasts, universally regarded as one world-leading institutes producing forecasts. wide rankings, depending region variable (with summer rainfall over Northern Europe performing exceptionally poorly) found. Finally, discuss prospects reaching all regions variables time.