Abstract
Forecasting severe weather events is of great importance, especially since it can have impact on life and society. Even though numerical weather prediction (NWP) models have had great improvements in the last decade, some severe cyclogenesis fail to be forecasted, especially in data-sparse regions. For meteorologists on duty it is important to use all available information to make the best possible forecast. Potential vorticity (PV) field can be superimposed onto a water vapour (WV) image to find mismatches, or similarities, between
the output coming from the NWP model and the reality. Comparison between pseudo water vapour (PWV) images and WV images can also be used for the same purpose.
This thesis investigates four selected low pressure developments during autumn 2009 and spring 2010, with emphasis on comparison between PV fields and WV images, or between WV images and PWV images. If deviations were found in the interpretation between these fields, an attempt to improve the initial state of the weather forecast was made with PV modifications.
In one of these cases there was mismatch between the PV field in the upper troposphere and the WV image, and also between the PWV image and
the WV image. In this case PV modification was done, and a new initial
height field was obtained by PV inversion. The new modified height field was compared with the ensemble members from NORLAMEPS (Norwegian
Limited Area Model Ensemble Prediction System). The ensemble members
which where similar to the new modified initial field were chosen. 54% of these members gave a better prognosis of the low pressure system than the original forecast, in comparison to 17.5% before using this method. On this basis a meteorologist on duty could have made a better forecast.