Adrian investigates causes of major poor forecasts with an emphasis on data assimilation.
Adrian is a senior scientist working on the identification and diagnosis of major forecast errors from the Unified Model. Forecasts from numerical weather prediction (NWP) models occasionally contain significant errors and investigation of cases can sometimes identify their causes. Adrian's work is aimed primarily at investigating observation and data assimilation issues, but model error not attributable to these aspects is investigated in collaboration with relevant research groups. The ultimate aim of Adrian's work is to improve the Met Office forecasting system. Effective diagnosis of a forecast error requires an overall understanding of the technical aspects of the forecasting process along with a knowledge of the meteorological processes which have occurred and Adrian uses error tracking techniques which incorporate both these aspects. Understanding of the processes involved in error development and propagation can then help identify the dominant error source with re-running of the Unified Model under changed conditions often necessary to investigate an error to conclusion. Not all investigations lead to an identifiable error-source, however, due both to the complexity and predictability issues of numerical weather prediction. The identification and diagnosis of forecast errors can come from routine monitoring, but can also come from close liaison with the Met Office's operational forecasters. In contrast, the application of diagnostic techniques developed in the course of post-event analysis has been found to be fruitful in the operational environment, aiding in the decision making process and improving the forecasting end-product. Adrian's work therefore provides a strong and important link between research and operational meteorology.
Adrian has been a member of DAE since joining the Met Office in June 1997. Prior to joining the Met Office, Adrian completed a Pure and Applied Physics Degree (first class) at Liverpool University, where he remained to complete a PhD in Nuclear Structure Physics and two years postdoctoral research.On the joining the Met Office, Adrian worked for five years at JCMM University of Reading where he worked on a number of projects including the application of conceptual models of cyclogenesis in NWP. Here he developed a unification of these models which has since been used widely by operational forecasters and to teach the meteorology of cyclogenesis. Adrian also developed code and studied the effects of an error-breeding system for the Unified Model which provided an ideal progression to his current work as forecast investigator.