Simulating Weather Forecast Errors in Traffic Flow Management Applications

Abstract

This paper describes a model and related algorithms for fast-time simulation of weather forecast errors in air traffic flow management applications. The model uses currently available convective weather datasets and generates forecasts using three types of error transformations: time-shift, location, and coverage. It produces a single deterministic convective weather forecast with fully-parameterized error bounds for a given look-ahead time. We illustrate the model with several examples of generated forecasts with predetermined errors, and discuss their applicability to a range of applications in traffic flow management in which weather uncertainty constitutes an important factor. We also present initial conclusions and recommendations for future development of the model. Nomenclature eC(t) = Magnitude of the coverage error transformation at time t eL,X(t) = Magnitude of the location error transformation in x direction at time t eL,Y(t) = Magnitude of the location error transformation in y direction at time t eT(t) = Magnitude of the time-shift error transformation at time t δgrid = Grid spacing for convective weather data product η(xi,yj,r) = Set of weather data points contained within a local neighborhood of radius r Δr = Time-shift error ΔL = Location error vector ΔC = Coverage error D = Spatial domain for which convective weather data are available I(xi,yj) = Intensity of convective weather at location (xi,yj) N = Number of available convective weather nowcasts r = Radius of the local neighborhood for defining the coverage error transformation t n Wx = Times at which weather nowcast data are available t k T = Times at which error transformation bounds are defined