1	Integrating Convective Wx into Air Traffic Management Decision Support Tools July 15, 2008 Robert Beard
2	Overview Background What information is needed? Assessment– How well are we doing today? Notional Steps for Integration of Convective Wx Open Issues/Needed Research Summary
3	Background– Dramatic rise in fuel prices, airline profitability concerns heighten demands for Increased accuracy of wx prediction and consequent flight scheduling efficiency Minimizing in-flight delays Efficient routing around storm cells Elimination of unnecessary re-routing Efficient, incremental recovery after storm has cleared Increased congestion makes responses to predicted wx impacts even more critical
4	What information is needed for Air Traffic Management? Accurate wx forecasts (increased precision, less uncertainty) Prediction of winds, icing, temps, pressure Prediction of convective cells: Predicted position and velocity Intensity, echo tops Shape Growth/decay predictions uncertainty Enhancements to ATM Decision Support Tools to accommodate advanced wx forecasts and uncertainties
5	Convective wx prediction today Convective Wx forecasting: Near-term [0-1 hr] (E.g.: TCWF, ITWS, CIWS) Long-term [2-6 hr] (E.g.: Collaborative Convective Forecast Product– large uncertainties Self-assessment of accuracy can prove a valuable accuracy metric Experimental collaborative convective product (Terminal) [2-5 hr] Need: 0-2 hrs for avg duration flights; 2-6 hrs for planning & for longer flights Well-defined “forced” cells more accurately predicted than “pop-up” convection
6	Integration of convective wx into today’s DSTs Most DSTs operational in today’s NAS: Make no attempt to model dynamic convective wx cells Rely on the controller/TMC to fuse data and tactically (manually) make traffic adjustments Lack automation to generate suggested traffic management responses to predicted convective cells Issues limiting progress Accuracy and precision of convective cell predictions, especially for 2+ hours Pop-up cells Software modifications/restructuring required for legacy DST Some exceptions are: RAPT Prototypes of TMA Other R&D prototypes
7	Notional steps for integration of convective wx 1. Display wx cell as an overlay on ATM Situation Display Real time and/or predicted Controller/TMC doesn’t have to mentally superimpose, extrapolate 2. Compute blocked airspace as function of time Initially as rigid simplistic volume of airspace (“moving SUA/FCA”) Subsequently refine for dynamic shape and intensity 3. Auto-generate response options to blocked airspace Determine alternative routing and/or configuration Reduce sector/route capacity; manage impact to near-by sectors Reallocate unused capacity from blockage to unaffected streams 4. Automate recovery planning for wx-impacted flights Restoration profile for capacities (e.g., per aircraft capabilities) Strategies for “draining” backlogged flows
8	Open Issues/Needed Research Establish/improve quantitative measures of performance: Prediction accuracy for cell position, velocity, intensity, shape, growth/decay, uncertainties Separation: Fully safe yet efficient separation between a/c and predicted convective wx as it forms, approaches, departs, dissipates Unused capacity: Automation systems to translate Wx predictions, system status, and user demand into efficient system operations Safety, efficiency operational metrics: normal èdegraded è recovery Develop/enhance algorithms to re-route and flow traffic around wx problems per these performance measures Establish/refine data repositories permitting “replay” of actual scenario results against enhanced algorithms Incorporate enhanced algorithms into tactical/strategic automation for traffic control and flow management Analyze/document new vs old operational procedures
9	Summary Much progress made in prediction accuracy for convective wx [0-1 hr] Need improved accuracies for longer term prediction to permit more adequate ATM planning and response Need enhancements to DSTs to accommodate advancements in convective wx predictions Where convective uncertainties are large, incorporate probability of occurrence into flow management planning/projections