Re-analysis of past weather

Atmospheric models can not only be used to create weather forecasts but also to give us a better understanding of past weather patterns. This can be achieved by a method called dynamical downscaling. The idea behind dynamical downscaling is relatively simple. Take output from a coarse resolution model, e.g. a Global Circulation Model (GCM), and use it to force a Limited Area Model (LAM) at a higher horizontal and vertical resolution. As resolution is increased, processes governed by the interaction of the large scale flow and topography become better resolved by the models. This is necessary as many impact assessment models require information at scales of 10 km or less, which is considerably more fine grained than the resolution of current GCMs.

Belgingur has to this day undertaken four downscaling projects where coarse resolution atmospheric reanalysis data from GCMs have been used as initial and boundary data to LAMs. The first of these was done back in 2003 where we used, at the time, the state-of-the-art PSU/NCAR MM5 atmospheric model, forced with data from the ERA40 reanalysis project. Since then, three other reanalysis data series have been created. The last one, named IceBox, spans from September 1990 to present, where we extend the simulations at regular intervals so that a typical delay is never more than three months.

Among other uses, the high-resolution data stemming from the reanalysis projects can be used as the first step towards wind resource assessments. This is done by calculating the potential wind power production at various heights above ground level for each grid cell in the model. From this a detailed wind atlas can be created for the region in question. Another use, closely related to the wind atlas, is the mapping of potential high-risk atmospheric icing regions. But icing on power lines can cause major disruptions in electricity supply networks.