The effect of targeted observations on the prediction of weather systems

Emma Irvine
Sue Gray
John Methven
Ian Renfrew (UEA)
Richard Swinbank (Met Office)

Targeted observations seek to reduce forecast busts which occur as a result of growth of initial condition error. Large areas of ocean and the high-latitudes lack detailed profile measurements (such as those provided by radiosondes) meaning that there are less observations available in these regions to constrain the background field there. This can lead to errors in the initial conditions in these regions, which may be small initially, that grow during the forecast, deteriorating the forecast for a region further downstream of this so-called 'sensitive' region. During the Greenland Flow Distortion Experiment (Renfrew et al, 2008) in Spring 2007 the Facility for Atmospheric Airborne Measurement aircraft was used to deploy dropsondes into sensitive regions around Iceland and southern Greenland. The impact of this additional observational data on the 1-2 day forecasts for northern Europe is being assessed with help from the Met Office. The properties of the data assimilation scheme used to assimilate the targeted observations may be important in determining the impact of the targeted observations on the forecast, and this is being investigated using the Met Office four-dimensional variational data assimilation scheme.

In Spring 2008 a developing polar low in the Norwegian seas was targeted during the Norwegian IPY-THORPEX field campaign (PI Jon Egill Kristjansson, Univ. Oslo) using both dropsondes deployed from, and LIDAR onboard, the DLR Falcon. The impact of this targeted data on the forecasts for Scandinavia is being investigated.

For more information follow these links:

• The Greenland Flow Distortion Experiment
• IPY-THORPEX
• Both experiments are part of the International Polar Year