There are three separate but closely related activities within the workpackage: Forward propagation model, Anomalous propagation identification, Development of a real-time application. This workpackage relates to NWP in two directions; namely, as an instrument in the improvement of radar (precipitation) data quality for input to NWP, and in using NWP to predict the incidence of ANAPROP. This approach, therefore, offers a new element in terms of the relation between radar data and modelling, namely synergy, in which weaknesses in each processes are complemented by strengths in the other.
Forward propagation model – Partner 6.
In a previous project ray tracing methods were used to identify correlation between ANAPROP conditions and radar images. Visualisation and quantitative prediction of the effects of ANAPROP requires more detailed modelling. Using a highly efficient parabolic equation technique, the distribution of illumination of sea or terrain can be determined. Because the full electromagnetic field is calculated, the propagation factors for the return paths are known, and from the reflectivity of the surface the contributions to the apparent ANAPROP radar image can be modelled. This defines a self-contained package capable of running locally on a PC at a radar site to map the model-predicted ANAPROP for the individual radar. The outputs would provide image products in real-time available to radar operators, and data to improve the performance of existing ANAPROP recognition and cancellation methodologies.
Anomalous propagation identification - (Partner 7)
Both super- and sub-refractive propagation conditions can significantly affect radar observations. This component will study the propagation conditions retrieved from radiosonde data – both high resolution and standard TEMP products, and from NWP mesoscale data, building a dataset of significant profiles. The outcome of the research envisaged would be the development and evaluation of ANAPROP diagnostic product, which might provide a flag for inclusion weather radar products, or as input to radar processing algorithms to enable refinement of identification and removal of ANAPROP effects. The experience and analysis of this element will provide important guiding factors in the development of the radar simulation applications of the workpackage.
Development of a real-time application (Partner 6)
The final element involves building on the first two. This is to combine the application of the forward propagation model with the assimilation of observed radar data to identify, and if possible correct the NWP predictions of refractivity. The radar image is very sensitive to heights and slopes of stable atmospheric layers, and such data are therefore potentially useful in constraining in turn the initialisation of NWP. This will require iterative evaluation of the propagation model combined with processing of the actual radar data, which merits a parallel approach. Recently, parallel processing using closely-coupled multiple processors and high-speed graphics and data bus has become available at modest cost (e.g. SGI 540 with 4 Pentium III Xeon processors) greatly reducing potential processing bottlenecks. We will develop a methodology for combining the simulation and data acquisition for display and data networking, resulting in a deliverable package, demonstrating on stored sequences of data, collected during the project. The final package should also incorporate the ANAPROP diagnostic product of the second component of the workpackage into the display, so integrating the entire workpackage into a demonstration and evaluation system.