A new weather forecast system for air traffic using real-time observations and model data

Arnold Tafferner and Caroline Forster

German Aerospace Center (DLR), Institute of Atmospheric Physics, Oberpfaffenhofen D-82230 Wessling, Germany
Email: caroline.forster@dlr.de , arnold.tafferner@dlr.de

Motivation and Goal

Recent studies identify weather as the primary reason for delays and disruptions in the air transport system. For instance, during the summer months flights are often delayed or even cannot be operated because of thunderstorms in the terminal manoeuvring area, while during winter operations at the airports might be substantially affected by winter storms, snowfall and icing conditions. There is a clear need for integrated systems that analyse and forecast weather hazards for air traffic as precisely as possible in order to enable the mitigation of the weather hazard's effects. Several such systems have been developed for U.S. airports during the last decade. In Europe, an important piece of the ACARE (Advisory Council for Aeronautics in Europe) plan has been put in place early in 2005: the FLYSAFE Project (http://www.eu-flysafe.org/). FLYSAFE aims at defining and testing new tools and systems contributing to the safety of flights for all aircraft. It focuses on the development of new on-board systems and of the tools on the ground for feeding them with the information that they require. Among them also weather plays a significant role. In accord with these efforts a new project has started in 2008 in Germany under the leadership of the Institute for Atmospheric Physics (IPA) at the German Aerospace Center (DLR) in Oberpfaffenhofen. Named "Weather and Flying", this DLR project aims at increasing safety and efficiency of air traffic and to secure the competitiveness of German and European aviation industry. In order to reach this high level goal two main systems are being constructed:

• Integrated airport weather systems for the airports Frankfurt and Munich comprising the components "wake cortices", "thunderstorms" and "winter weather"
• On board and ground based systems for steering, monitoring and information in order to improve flying characteristics in case of turbulence, wake vortices and thunderstorms.

This article sheds light on the integrated weather systems part of this project by introducing the Weather Forecast User Oriented System Including Object Nowcasting (WxFUSION). This system is envisaged to be operationally installed at the airports Frankfurt and Munich during the Weather and Flying project and is build using IDL programming and development tools.

The Concept of the System

Figure 1: Schematic diagram of the WxFUSION concept. User specified target weather objects (TWO) are extracted by fusion of selected observations and nowcasting information (upper half) and forecast products (lower half).

WxFUSION aims at combining real-time observations from different data sources with nowcasting tools and numerical model simulations (Fig. 1). This combination has the benefit that the assertions of the individual tools, e.g. with regard to the exact location of a particular weather system, its intensity and further development, can be processed and contrasted. Thus, the system can be expected to provide a more reliable interpretation of the actual and future state of a weather system than only one data source would be able to provide. The system's core element "FUSION" will combine available data from the various tools accordingly in order to detect, nowcast (0-1 hrs), and forecast (1-24 hrs) user-defined target weather objects (TWO) like thunderstorms and winter weather. The fusion process will be based on fuzzy logic which allows to account for imprecise observations and forecasts and also to deal with parameter ranges instead of fixed thresholds. Furthermore conceptual models, expert knowledge can be incorporated this way. In addition, local constraints and requirements for safe aircraft operations with respect to certain weather conditions can be accounted for novel tools like synthetic satellite (SYNSAT) and radar images (SYNRAD, SYNPOLRAD) together with an image matching tool developed at DLR allow estimating the forecast quality of local high resolution as well as large-scale ensemble weather forecasts ("Forecast Control"). In particular, these tools allow selecting the numerical model forecast that agrees best with the observations and can be used to extend the nowcasting horizon (0-1 hour) with forecasting for several hours (1-6).

System Realisation by Use of IDL Functionality and Graphical User Interfaces

Figure 2 (click to enlarge): WxFUSION IDL GUI showing satellite image in the visible channel and thunderstorm objects marked by red (mature thunderstorm) and orange (developing) over southern Germany and the Northern Alps on 12 August 2004. Also given are forecast positions of the objects (grey contours). Weather attributes of the objects can be visualised with mouse-over display. Additional buttons allow the user to select the case, to navigate in time, to choose among data sources (output from various nowcasting tools and forecasting models) and the objects to be displayed.

For integration of data from different sources into WxFUSION, IDL has been chosen for mainly two reasons. Firstly, the existing nowcasting tools have been already coded in IDL. E.g. the development of the cloud tracker CB TRAM has been carried out in IDL due to its efficient array handling and structure components allowing these clouds to be described by their geometrical properties as well as by weather attributes. Secondly, IDL provides in ideal platform during the development of WxFUSION, where algorithms are changed frequently and the effect of the changes has to be analysed quickly. For the latter, a graphical user interface has been constructed in IDL, not only for the overlay of the different weather observations with the results of the nowcasting tools and the numerical forecasts, but also for making visible the result of the data fusion process. The system is set up in a way that it can be adjusted to local constraints at different terminal manoeuvring areas. Figure 2, a screenshot of the WXFUSION GUI, shows a large thunderstorm cell in the visible satellite image (grey shading) over the area of Munich, Germany, on 12 August 2004 at 17 UTC. The dark red contour marks a thunderstorm TWO as identified by the cloud tracker Cb-TRAM, orange contours mark new development cells and grey contours are nowcast positions of the cells 60 minutes in advance. The users (in future e.g. weather providers and flight controllers) can navigate in time and identify and track the TWOs. In order to provide short, clear and precise information for quick decision making, the TWOs are described as idealized objects with individual weather attributes like moving speed, moving direction, lightning density, developing stage (growing/decaying), and severity level (moderate/severe). This additional information can be visualized e.g. by mouse over displays as shown in the figure. During the FLYSAFE project a database with weather situations has already been put in place. The tools and algorithms integrated in WxFUSION can now be adjusted with these data and optimized to the conditions at the airports of Frankfurt and Munich. For this the development environment built on IDL provides the necessary framework.