Abstract/Summary

An atmospheric block is a weather event characterised by a large-amplitude tropospheric ridge of high pressure that is synoptic in scale and can persist over a region for several weeks. Such blocks, occur frequently in the midlatitudes of both hemispheres and can be associated with high-impact weather. The prediction of blocking events has been a well-known and frequently-studied problem in both weather and climate models for many decades. Medium-range forecasting of blocking in the northern hemisphere is assessed here, with a focus on improving the predictability of blocking originating from the representation of upstream cyclones and the impacts of improvements to numerical weather prediction models. An improvement to the dynamical core of the Met Office Unified Model (MetUM) is proven to counteract the consistent underprediction of block frequency by weather and climate models. Implementing planned Met Office upgrades to various physical parameterisation schemes, and running the MetUM with updated sea surface temperatures (SST), are also shown to impact forecast evolution in a case study of upper-level ridge amplification and block development. The sensitivity of block development in the forecast results from different diabatic heating rates from parameterised processes: stronger diabatic heating in the warm conveyor belt of an extratropical cyclone led to a moreamplified ridge. Using updated SST does not distinguishably impact forecast skill when averaged over 54 forecasts initialised during autumn 2016, whereas using an improved representation of convection reduces forecast error. Accurate prediction of the intensities and locations of extratropical cyclones that feed air into blocking ridges via their warm conveyor belts is shown to be important for the most unpredictable blocking forecasts from the ECMWF. Together this work implies that improvements to weather models have the potential to further improve forecasts of blocking events which can cause of some of the worst forecasts produced at forecasting centres.